We propose a method for generating spurious features by leveraging large-scale text-to-image diffusion models. Although the previous work detects spurious features in a large-scale dataset like ImageNet and introduces Spurious ImageNet, we found that not all spurious images are spurious across different classifiers because images that contain spurious features were filtered by a limited number of classifiers. Although spurious images help measure the reliance of a classifier, filtering many images from the Internet to find more spurious features is time-consuming. To this end, we utilize an existing approach of personalizing large-scale text-to-image diffusion models with available discovered spurious images and propose a new spurious feature similarity loss based on neural features of an adversarially robust model. Precisely, we fine-tune Stable Diffusion with several reference images from Spurious ImageNet with a modified objective incorporating the proposed spurious-feature similarity loss. Experiment results show that our method can generate spurious images that are consistently spurious across different classifiers. Moreover, the generated spurious images are visually similar to reference images from Spurious ImageNet.

In this paper, we propose a speech pseudonymization framework that utilizes cascaded and superposition-based voice modification modules. With increasing opportunities to use spoken dialogue systems nowadays, research regarding protecting the privacy of speaker information encapsulated in speech data is attracting attention. Pseudonymization, which is one method for voice privacy protection, aims to keep the intelligibility of speech while simultaneously suppressing speaker-specific information. One motivation of our framework is to achieve a reliable pseudonymization performance with light computation. To do this, we utilize the advantages of both machine learning-based and signal processing-based approaches. The advantages are (1) using signal processing-based methods parameterized with few hyperparameters and (2) using machine learning-based optimization to optimize all hyperparameters on the basis of black-box systems consisting of automatic speaker verification and automatic speech recognition. Our method of cascading signal processing modules, which are jointly optimized in a data-driven manner, can pseudonymize speech in a lightweight manner. Additionally, we discuss irreversible pseudonymization approaches and propose a superposition approach, yet another pseudonymization method that is more irreversible than the cascade method in terms of estimating the adequate parameters to recover the original signal. From the experimental results conducted under the VoicePrivacy 2020 protocols, we can demonstrate that (1) our cascade method succeeds in deteriorating the speaker recognition rate by over 24% while simultaneously improving the speech recognition rate by approximately 8% compared with a signal processing-based baseline system of VoicePrivacy 2020 and that (2) our superposition method works comparable to our cascade method in terms of pseudonymization performance.

In this paper, we propose a novel multi-color balance adjustment for color constancy. The proposed method, called n-color balancing allows us not only to perfectly correct n target colors on the basis of corresponding ground truth colors but also to correct colors other than the n colors. In contrast, although white-balancing can perfectly adjust white, colors other than white are not considered in the framework of white-balancing in general. In an experiment, the proposed multi-color balancing is demonstrated to outperform both conventional white and multi-color balance adjustments including Bradford’s model.

We propose a novel method for protecting trained models with a secret key so that unauthorized users without the correct key cannot get the correct inference. By taking advantage of transfer learning, the proposed method enables us to train a large protected model like a model trained with ImageNet by using a small subset of a training dataset. It utilizes a learnable encryption step with a secret key to generate learnable transformed images. Models with pre-trained weights are fine-tuned by using such transformed images. In experiments with the ImageNet dataset, it is shown that the performance of a protected model was close to that of a non-protected model when the correct key was given, while the accuracy tremendously dropped when an incorrect key was used. The protected model was also demonstrated to be robust against key estimation attacks.

In this paper, we propose a novel method for separately estimating spectral distributions from images captured by a typical RGB camera. The proposed method allows us to separately estimate a spectral distribution of illumination, reflectance, or camera sensitivity, while recent hyperspectral cameras are limited to capturing a joint spectral distribution from a scene. In addition, the use of Bayesian inference makes it possible to take into account prior information of both spectral distributions and image noise as probability distributions. As a result, the proposed method can estimate spectral distributions in a unified way, and it can enhance the robustness of the estimation against noise, which conventional spectral-distribution estimation methods cannot. The use of Bayesian inference also enables us to obtain the confidence of estimation results. In an experiment, the proposed method is shown not only to outperform conventional estimation methods in terms of RMSE but also to be robust against noise.

In this paper, we propose a novel DNN watermarking method that utilizes a learnable image transformation method with a secret key. The proposed method embeds a watermark pattern in a model by using learnable transformed images and allows us to remotely verify the ownership of the model. As a result, it is piracy-resistant, so the original watermark cannot be overwritten by a pirated watermark, and adding a new watermark decreases the model accuracy unlike most of the existing DNN watermarking methods. In addition, it does not require a special pre-defined training set or trigger set. We empirically evaluated the proposed method on the CIFAR-10 dataset. The results show that it was resilient against fine-tuning and pruning attacks while maintaining a high watermark-detection accuracy.

In this paper, we propose a voice anonymization framework based on data-driven optimization of cascaded voice modification modules. With increasing opportunities to use speech dialogue with machines nowadays, research regarding privacy protection of speaker information encapsulated in speech data is attracting attention. Anonymization, which is one of the methods for privacy protection, is based on signal processing manners, and the other one based on machine learning ones. Both approaches have a trade off between intelligibility of speech and degree of anonymization. The proposed voice anonymization framework utilizes advantages of machine learning and signal processing-based approaches to find the optimized trade off between the two. We use signal processing methods with training data for optimizing hyperparameters in a data-driven manner. The speech is modified using cascaded lightweight signal processing methods and then evaluated using black-box ASR and ASV, respectively. Our proposed method succeeded in deteriorating the speaker recognition rate by approximately 22% while simultaneously improved the speech recognition rate by over 3% compared to a signal processing-based conventional method.

In this paper, we propose a novel defensive transformation that enables us to maintain a high classification accuracy under the use of both clean images and adversarial examples for adversarially robust defense. The proposed transformation is a block-wise preprocessing technique with a secret key to input images. The proposed defense obfuscates gradients in the absence of the secret key unlike previously defeated obfuscating defenses. We developed three algorithms to realize the proposed transformation: Pixel Shuffling, Bit Flipping, and FFX Encryption. Experiments were carried out on the CIFAR-10 and ImageNet datasets by using both black-box and white-box attacks with various metrics including adaptive ones. The results show that the proposed defense achieves high accuracy close to that of using clean images even under adaptive attacks for the first time. In the best-case scenario, a model trained by using images transformed by FFX Encryption (block size of 4) yielded an accuracy of 92.30% on clean images and 91.48% under PGD attack with a noise distance of 8/255, which is close to the non-robust accuracy (95.45%) for the CIFAR-10 dataset, and it yielded an accuracy of 72.18% on clean images and 71.43% under the same attack, which is also close to the standard accuracy (73.70%) for the ImageNet dataset. Overall, all three proposed algorithms are demonstrated to outperform state-of-the-art defenses including adversarial training whether or not a model is under attack.

Big-data analysis by edge/cloud systems is becoming more important. However, when information may lead to personal identification, such information tends to be encrypted and restricted to its owners to ensure privacy protection. The resulting data is often insufficiently detailed to permit useful analysis. As a result, the desired analysis accuracy may not be achieved. To deal with this issue, several studies have examined encryptions based on the random unitary transform. This is because the random unitary transform has lower computational complexity than other encryption schemes, and its encryption domain supports several signal processing algorithms. However, analysis models on distributed encrypted data, have not been studied deeply enough. In this paper, we construct an analysis model for data encrypted with the random unitary transform by deriving a LASSO solution for encrypted data. The analytical model can derive the same LASSO solution as that yielded by processing the original data (i.e. without encryption). The analytical model supports distributed encryption, where a data set consists of different components that are encrypted at different sites independently. The collaboration enables us to improve the accuracy of analysis for distributed privacy-sensitive information.

In this paper, we propose a fixed convolutional layer with an order of smoothness not only for avoiding checkerboard artifacts in convolutional neural networks (CNNs) but also for enhancing the performance of CNNs, where the smoothness of its filter kernel can be controlled by a parameter. It is well-known that a number of CNNs generate checkerboard artifacts in both of two process: forward-propagation of upsampling layers and backward-propagation of strided convolutional layers. The proposed layer can perfectly prevent checkerboard artifacts caused by strided convolutional layers or upsampling layers including transposed convolutional layers. In an image-classification experiment with four CNNs: a simple CNN, VGG8, ResNet-18, and ResNet-101, applying the fixed layers to these CNNs is shown to improve the classification performance of all CNNs. In addition, the fixed layer are applied to generative adversarial networks (GANs), for the first time. From image-generation results, a smoother fixed convolutional layer is demonstrated to enable us to improve the quality of images generated with GANs.

In this paper, we propose a privacy-preserving pattern recognition method that uses encrypted sparse representations in L0 norm minimization. We prove, theoretically, that the proposal has exactly the same dictionary and sparse coefficient estimation performance as the Label Consistent K-Singular Value Decomposition (LC-KSVD) algorithm for non-encrypted signals. It can be directly implemented by the LC-KSVD algorithm without any modification. Finally, we demonstrate its excellent recognition performance and security strength for the face recognition task using the Extended YaleB database.

Conventional adversarial defenses reduce classification accuracy whether or not a model is under attacks. Moreover, most of image processing based defenses are defeated due to the problem of obfuscated gradients. In this paper, we propose a new adversarial defense which is a defensive transform for both training and test images inspired by perceptual image encryption methods. The proposed method utilizes a block-wise pixel shuffling method with a secret key. The experiments are carried out on both adaptive and non-adaptive maximum-norm bounded white-box attacks while considering obfuscated gradients. The results show that the proposed defense achieves high accuracy (91.55%) on clean images and (89.66%) on adversarial examples with noise distance of 8/255 on CFAR-10 dataset. Thus, the proposed defense outperforms state-of-the-art adversarial defenses including latent adversarial training, adversarial training and thermometer encoding.

We propose an image identification scheme for double-compressed encrypted JPEG images that aims to identify encrypted JPEG images that are generated from an original JPEG image. To store images without any visual sensitive information on photo sharing services, encrypted JPEG images are generated by using a block-scrambling-based encryption method that has been proposed for Encryption-then-Compression systems with JPEG compression. In addition, feature vectors robust against JPEG compression are extracted from encrypted JPEG images. The use of the image encryption and feature vectors allows us to identify encrypted images recompressed multiple times. Moreover, the proposed scheme is designed to identify images re-encrypted with different keys. The results of a simulation show that the identification performance of the scheme is high even when images are recompressed and re-encrypted.

In this paper, we propose a novel convolutional neural network (CNN) architecture considering both local and global features for image enhancement. Most conventional image enhancement methods, including Retinex-based methods, cannot restore lost pixel values caused by clipping and quantizing. CNN-based methods have recently been proposed to solve the problem, but they still have a limited performance due to network architectures not handling global features. To handle both local and global features, the proposed architecture consists of three networks: a local encoder, a global encoder, and a decoder. In addition, high dynamic range (HDR) images are used for generating training data for our networks. The use of HDR images makes it possible to train CNNs with better-quality images than images directly captured with cameras. Experimental results show that the proposed method can produce higher-quality images than conventional image enhancement methods including CNN-based methods, in terms of various objective quality metrics: TMQI, entropy, NIQE, and BRISQUE.

We propose an efficient two-layer near-lossless coding method using an extended histogram packing technique with backward compatibility to the legacy JPEG standard. The JPEG XT, which is the international standard to compress HDR images, adopts a two-layer coding method for backward compatibility to the legacy JPEG standard. However, there are two problems with this two-layer coding method. One is that it does not exhibit better near-lossless performance than other methods for HDR image compression with single-layer structure. The other problem is that the determining the appropriate values of the coding parameters may be required for each input image to achieve good compression performance of near-lossless compression with the two-layer coding method of the JPEG XT. To solve these problems, we focus on a histogram-packing technique that takes into account the histogram sparseness of HDR images. We used zero-skip quantization, which is an extension of the histogram-packing technique proposed for lossless coding, for implementing the proposed near-lossless coding method. The experimental results indicate that the proposed method exhibits not only a better near-lossless compression performance than that of the two-layer coding method of the JPEG XT, but also there are no issue regarding the combination of parameter values without losing backward compatibility to the JPEG standard.

We present a novel privacy-preserving scheme for deep neural networks (DNNs) that enables us not to only apply images without visual information to DNNs for both training and testing but to also consider data augmentation in the encrypted domain for the first time. In this paper, a novel pixel-based image encryption method is first proposed for privacy-preserving DNNs. In addition, a novel adaptation network is considered that reduces the influence of image encryption. In an experiment, the proposed method is applied to a well-known network, ResNet-18, for image classification. The experimental results demonstrate that conventional privacy-preserving machine learning methods including the state-of-the-arts cannot be applied to data augmentation in the encrypted domain and that the proposed method outperforms them in terms of classification accuracy.

A block scrambling-based encryption scheme is presented to enhance the security of Encryption-then-Compression (EtC) systems with JPEG compression, which allow us to securely transmit the images through an untrusted channel provider, such as social network service providers. The proposed scheme enables the use of a smaller block size and a larger number of blocks than the conventional scheme. Images encrypted using the proposed scheme include less color information due to the use of grayscale images even when the original image has three color channels. These features enhance security against various attacks such as jigsaw puzzle solver and brute-force attacks. In an experiment, the security against jigsaw puzzle solver attacks is evaluated. Encrypted images were uploaded to and then downloaded from Facebook and Twitter, and the results demonstrated that the proposed scheme is effective for EtC systems.

We propose a novel method for adjusting luminance for multi-exposure image fusion. For the adjustment, two novel scene segmentation approaches based on luminance distribution are also proposed. Multi-exposure image fusion is a method for producing images that are expected to be more informative and perceptually appealing than any of the input ones, by directly fusing photos taken with different exposures. However, existing fusion methods often produce unclear fused images when input images do not have a sufficient number of different exposure levels. In this paper, we point out that adjusting the luminance of input images makes it possible to improve the quality of the final fused images. This insight is the basis of the proposed method. The proposed method enables us to produce high-quality images, even when undesirable inputs are given. Visual comparison results show that the proposed method can produce images that clearly represent a whole scene. In addition, multi-exposure image fusion with the proposed method outperforms state-of-the-art fusion methods in terms of MEF-SSIM, discrete entropy, tone mapped image quality index, and statistical naturalness.

This paper proposes a novel multi -exposure image fusion method based on exposure compensation. Multi-exposure image fusion is a method to produce images without color saturation regions, by using photos with different exposures. However, in conventional works, it is unclear how to determine appropriate exposure values, and moreover, it is difficult to set appropriate exposure values at the time of photographing due to time constraints. In the proposed method, the luminance of the input multi-exposure images is adjusted on the basis of the relationship between exposure values and pixel values, where the relationship is obtained by assuming that a digital camera has a linear response function. The use of a local contrast enhancement method is also considered to improve input multi-exposure images. The compensated images are finally combined by one of existing multi-exposure image fusion methods. In some experiments, the effectiveness of the proposed method are evaluated in terms of the tone mapped image quality index, statistical naturalness, and discrete entropy, by comparing the proposed one with conventional ones.

This paper proposes a novel luminance adjustment method based on automatic exposure compensation for multi -exposure image fusion. Multi-exposure image fusion is a method to produce images without saturation regions, by using photos with different exposures. In conventional works, it has been pointed out that the quality of those multi-exposure images can be improved by adjusting the luminance of them. However, how to determine the degree of adjustment has never been discussed. This paper therefore proposes a way to automatically determines the degree on the basis of the luminance distribution of input multi -exposure images. Moreover, new weights, called “simple weights”, for image fusion are also considered for the proposed luminance adjustment method. Experimental results show that the multi -exposure images adjusted by the proposed method have better quality than the input multi-exposure ones in terms of the well-exposedness. It is also confirmed that the proposed simple weights provide the highest score of statistical naturalness and discrete entropy in all fusion methods.

It is well-known that a number of excellent super-resolution (SR) methods using convolutional neural networks (CNNs) generate checkerboard artifacts. A condition to avoid the checkerboard artifacts is proposed in this paper. So far, checkerboard artifacts have been mainly studied for linear multirate systems, but the condition to avoid checkerboard artifacts can not be applied to CNNs due to the non-linearity of CNNs. We extend the avoiding condition for CNNs, and apply the proposed structure to some typical SR methods to confirm the effectiveness of the new scheme. Experiment results demonstrate that the proposed structure can perfectly avoid to generate checkerboard artifacts under two loss conditions: mean square error and perceptual loss, while keeping excellent properties that the SR methods have.

This paper proposes a new block scrambling encryption scheme that enhances the security of encryption-then-compression (EtC) systems for JPEG images, which are used, for example, to securely transmit images through an untrusted channel provider. The proposed method allows the use of a smaller block size and a larger number of blocks than the conventional ones. Moreover, images encrypted using proposed scheme include less color information due to the use of grayscale even when the original image has three color channels. These features enhance security against various attacks such as jigsaw puzzle solver and brute-force attacks. The results of an experiment in which encrypted images were uploaded to and then downloaded from Twitter and Facebook demonstrated the effectiveness of the proposed scheme for EtC systems.

A novel method using the histogram packing technique with the two-layer coding having the backward compatibility to the legacy JPEG for base layer is proposed in this paper. The histogram sparseness of HDR images is discussed and it is pointed out that the histogram packing technique considering the sparseness is able to improve the performance of lossless compression for HDR images. The experimental results demonstrate that not only the proposed method has a higher compression performance than that of the JPEG XT part 8, but also there is no need to determine image-dependent parameter values for good compression performance. Moreover, the base layer produced by the proposed method has the backward compatibility to the well known legacy JPEG standard.

An efficient two-layer coding method using the histogram packing technique with the backward compatibility to the legacy JPEG is proposed in this paper. The JPEG XT, which is the international standard to compress HDR images, adopts two-layer coding scheme for backward compatibility to the legacy JPEG. However, this two-layer coding structure does not give better lossless performance than the other existing single-layer coding methods for HDR images. Moreover, the JPEG XT has problems on determination of the lossless coding parameters; Finding appropriate combination of the parameter values is necessary to achieve good lossless performance. The histogram sparseness of HDR images is discussed and it is pointed out that the histogram packing technique considering the sparseness is able to improve the performance of lossless compression for HDR images and a novel two-layer coding with the histogram packing technique is proposed. The experimental results demonstrate that not only the proposed method has a better lossless compression performance than that of the JPEG XT, but also there is no need to determine image-dependent parameter values for good compression performance in spite of having the backward compatibility to the well known legacy JPEG standard.

This paper is among the first to adapt automatic jigsaw puzzle solvers, which are methods of assembling jigsaw puzzles, to the field of information security. Block-based scrambling encryption schemes, which have been mainly proposed for Encryption-then-Compression (ETC) systems, have enough key spaces for protecting brute-force attacks. However, each block in encrypted images has almost the same correlation as that of original images. Therefore, it is required to consider the security from different viewpoints from number theory-based encryption methods with provable security such as RSA and DES. In this paper, existing jigsaw puzzle solvers are first reviewed in terms of attacking strategies on encrypted images. Then, a new jigsaw puzzle solver is proposed to extend some limitations of the conventional ones. In the experiments, the jigsaw puzzle solvers are applied to encrypted images to consider the security conditions of the encryption schemes.

This paper proposes a novel inverse TMO, which enables to generate HDR images from LDR ones, not only without using any specific parameters but also at low computing costs. Furthermore, the inverse TMO has a new characteristic when an LDR image is mapped from an HDR one by Reinhard's global operator. In the case, the HDR image reconstructed by the proposed method without parameters can be remapped into the same image as that remapped from an HDR one reconstructed with parameters. Experimental results show that the proposed inverse tone mapping (TM) operation can be carried out, while keeping better structural similarity and lower computing cost than conventional methods.

A robust identification scheme for JPEG images is proposed in this paper. The aim is to robustly identify JPEG images generated from the same original image, under various compression conditions such as differences in compression ratios and initial quantization matrices. The proposed scheme does not in principle produce any false negative matches. Features without visual information are used to achieve not only a robust identification scheme but also secure one. Conventional schemes can produce false negative matches under certain compression conditions and require the use of a secret key for secure identification. The proposed scheme is well-suited for the uploading images to social network services and for image retrieval and forensics. Simulations demonstrated the effectiveness of the proposed scheme outperformed conventional ones in terms of query performance, while maintaining reasonable security.

The aim of this paper is to adapt automatic jigsaw puzzle solvers, which are methods of assembling jigsaw puzzles, to the field of information security. Encryption-then-Compression (ETC) systems have been considered for the user-controllable privacy protection of digital images in social network services. Block scrambling-based encryption schemes, which have been proposed to construct ETC systems, have enough key spaces for protecting brute-force attacks. However, each block in encrypted images has almost the same correlation as that of original images. Therefore, it is required to consider the security from different viewpoints from number theory-based encryption methods with provable security such as RSA and DES. In this paper, existing jigsaw puzzle solvers are first reviewed in terms of attacking strategies on encrypted images. Then, an extended jigsaw puzzle solver is considered to improve some limitations of the conventional ones. In the experiments, the jigsaw puzzle solvers are applied to encrypted images to consider the security conditions of the encryption schemes.

The coding performance of the normative encoder of the JPEG XT in profile is analyzed and the problem on the encoder is summarized in this paper. It is pointed out that there is a mismatch in the handling of the quantization error between the normative encoder and the standard decoder. To avoid this problem, an improved structure has been proposed with consideration of the mismatch. The experimental results for the quality of the reconstructed HDR images demonstrate that the proposed structure may give a better coding performance than the normative encoder.

The super resolution technique plays an important role in displaying a conventional low resolution (LR) image on a recently developed high resolution (HR) display. In its process, some of enhancement techniques such as de-noising, de-blurring and tone-mapping are combined to increase visual quality. This paper introduces a subset noise bias compensation (NBC) to an LR-to-HR up-conversion process which is composed of tone-mapping and up-scaling. In our system, 1) an LR image is encoded and decoded with JPEG, 2) pixels in the decoded noisy image are tone-mapped, 3) observed pixel values are classifies into subsets and the bias of each subset is compensated with an optimally designed compensation value, and 4) the images are up-scaled with a state-of-the-art super-resolution technique. Improvement by the proposed method is confirmed with PSNR and SSIM evaluation measures, respectively.

Implementation issues of the separable 2D lifting wavelet transform has been discussed under the international standard JPEG 2000 scheme. Recently, the minimum lifting step structure of the transform was proposed introducing the multi-dimensional memory accessing. It can reduce latency in a parallel processing platform. However, its output signal is distorted due to truncation of multiplier coefficient values. Especially, a few coefficients are extremely sensitive to the truncation in this structure. Therefore a special treatment is required to these coefficients. In this paper, a tolerable (the maximum) truncation error is optimally assigned to each coefficient under a given cost function introducing the noise gain. It is applied to expressing coefficient values in the sum-of-power-of-two (SPT) format for reducing computational cost of multipliers of the transform.

A new Encryption-then-Compression (ETC) system for the JPEG 2000 standard is proposed in this paper. An ETC system is known as a system that makes image communication secure and efficient by using perceptual encryption and image compression. The proposed system uses the sign-scrambling and block-shuffling of discrete wavelet transform (DWT) coefficients as perceptual encryption. Unlike conventional ETC systems, the proposed system is compatible with the JPEG 2000 standard because the perceptually encrypted coefficients can be efficiently compressed by the JPEG 2000. The experimental results demonstrated that the proposed system achieved both acceptable compression performance and enough key-space for secure image communication while remaining compatible with the JPEG 2000 standard.

In many multimedia applications, image encryption has to be conducted prior to image compression. This paper proposes an Encryption-then-Compression system using a JPEG-friendly perceptual encryption method, which enables to be conducted prior to JPEG compression. The proposed encryption method can provides approximately the same compression performance as that of JPEG compression without any encryption, where both gray scale images and color ones are considered. It is also shown that the proposed system consists of four block-based encryption steps, and provide a reasonably high level of security. Most of conventional perceptual encryption methods have not been designed for international compression standards, but this paper focuses on applying the JPEG standard, as one of the most widely used image compression standards.

This paper proposes a de-noising method especially for a biased noise. Even though a noise is assumed to have zero mean, an image processing such as tone mapping alters this property. Firstly, a precise estimation of the noise bias caused by a tone mapping function is derived considering discontinuity of a general form of the tone mapping. Secondly, a method of compensating the noise bias is introduced investigating estimation from observed noisy pixel values. Thirdly, the derived estimation is applied not only to compensate the noise bias, but also to reduce variance of the biased noise. Finally, significance of the proposed method is evaluated using HDR images in combination with a non-local mean filter.

This paper proposes a cheat-prevention method of visual cryptography (VC) for binary images. A VC technique encrypts a secret image into images referred to as shares so that stacking sufficient shares recovers the secret image, where shares are kept by different parties. The proposed method prevents malicious parties from deceiving an honest party, i.e., from cheating the honest party. To fight cheating, no further pixel expansion is introduced to the proposed method, so this method improves the contrast of recovered secret image and cheat-prevention functionality.

We propose an integer tone mapping operator (TMO) for high dynamic range (HDR) images expressed in a floating-point data format. Two purposes are achieved by the proposed TMO. The first purpose is to implement a TMO with less memory space. The second purpose is to give an important step to realize a fixed-point TMO. The proposed TMO is available for HDR images in the OpenEXR format. The OpenEXR format has two numerical representations (the normalized number and the denormalized number) which are not in other HDR formats such as RGBE. These two numerical representations cause a problem in applying an integer TMO. The proposed method enables us to avoid the problem by using the intermediate format. Moreover, the exponent part and the mantissa part are processed separately as two integer numbers. As a result, an integer TMO with less numerical range is achieved by our method. The experimental results show that the proposed method can generate high-quality low dynamic range (LDR) images with less memory space.

One category of fast full-search block matching algorithms (BMAs) is based on the fast Fourier transformation (FFT). This paper proposes a parallel implementation method of FFT-based full-search BMAs. The FFT-based full-search BMAs are much faster than the direct full-search BMA, and its accuracy is as same as the direct full-search BMA. However, these are not designed for parallel processing. The proposed method divides the search window into multiple sub search windows using the overlap-save method, and the FFT-based full-search BMA is applied to each sub search window. These sub search windows are processed in parallel. By dividing the search window, the method can not only process in parallel, but also select the efficient FFT size. Furthermore, the method can also calculate two cross-correlations at the same time. These properties also contribute to speeding up of the block matching. The experimental results shows that the method on 6 cores CPU is about 11 times faster than the conventional FFT-based full-search BMA.

This report proposes a two layer lossless coding system for high dynamic range (HDR) images expressed in a floating point data format. Its encoder outputs compressed image data in two layers. From the base layer, a standard low dynamic range (LDR) image is decoded. Merging with the enhancement layer, the original HDR image is decoded without any loss. We introduce a reversible logarithmic mapping to reduce bit depth of pixel values. We also introduce bit plane classification to divide data into two layers. It was confirmed that our method has better coding performance than existing methods especially at high bit rate coding of LDR images.

This report proposes a new tone mapping operation which is implemented in integer input and integer output. A tone mapping operation (TMO) generates a low dynamic range (LDR) image from a high dynamic range (HDR) image. Since pixel values of an HDR image are generally expressed in a floating point data format, e.g. RGBE, OpenEXR, a TMO is also implemented in floating point calculations in conventional approaches. However, it requires huge memory resources, even though a resulting LDR image is expressed in simple integer. We perform a TMO with integer input and integer output to reduce memory resources. It is experimentally confirmed with PSNR and contrast evaluations that the proposed method offers LDR images of visually high quality comparable to the conventional method.

This report reduces the total number of lifting steps in a three-dimensional (3D) double lifting discrete wavelet transform (DWT), which has been widely applied for analyzing volumetric medical images. The lifting steps are necessary components in a DWT. Since calculation in a lifting step must wait for a result of former step, cascading many lifting steps brings about increase of delay from input to output. We decrease the total number of lifting steps introducing 3D memory accessing for the implementation of low delay 3D DWT. We also maintain compatibility with the conventional 5/3 DWT defined by JPEG 2000 international standard for utilization of its software and hardware resources. Finally, the total number of lifting steps and rounding operations were reduced to 67 % and 33 %, respectively. It was observed that total amount of errors due to rounding operations in the lifting steps was also reduced.

We propose a new technique of non-iterative super-resolution image reconstruction. A closed form solution to reconstruction is derived in the discrete cosine transform (DCT) domain from a MAP-based cost functional. An average image is used in order to avoid iterative operations. Symmetric convolution with appropriate types of DCT suppresses boundary distortion. Experimental results demonstrate the effectiveness of the proposed technique.

In this paper, a lossy data compression for a sparse histogram image signal is proposed. It is extended from an existing lossless coding which is based on a lossless histogram packing and a lossless coding. We introduce a lossy mapping, which has less computational load than the rate-distortion optimized Lloyd-Max quantization, and combine it with a lossless coding. It was confirmed that the proposed method attains higher performance in the rate-distortion plane than existing methods. This is because it can utilize histogram sparseness of images, and also its inverse mapping does not magnify quantization noise.

A hash based identification method of JPEG 2000 images with different coding parameters is presented. The method correctly identifies the compressed JPEG 2000 images encoded even if there are difference in the coding parameters, such as coding-rate and quantization step sizes. A new algorithm is introduced to calculate hash values from the number of zero-bit-planes that can be extracted from the JPEG 2000 codestream by only parsing the header information. Because the proposed approach uses the hash values, the length of data to identify images is fixed. Using the fixed length data, processing speed of the proposed approach should be fast. This feature is not provided by conventional methods. The experimental results revealed the effectiveness of the processing speed of the proposed method.

This paper proposes a new method of visually-lossless data hiding. The proposed method firstly applies a nonlinear quantization to an original image accordingly to the size of a payload to be hidden. The method, then, embeds the payload to the quantized image based on a histogram modification-based lossless data hiding technique. The hidden payload can be extracted reversibly and be completely removed from the image to recover the quantized image. Experimental results show the effectiveness of the proposed method.

In this report, we propose a lossy compression algorithm which can utilize the `histogram sparseness' of pixel values of an input image signal. The bit depth of pixel values has becoming longer than conventional ones in high dynamic range images. Their histogram of pixel values tends to be `sparse', - not all the histogram bins are used. We extend an existing lossless coding based on the histogram packing to a lossy coding, introducing a non-uniform quantization especially for histogram sparse images. It was confirmed that the proposed method attains finer control of the bit rate and also higher SNR in the rate-distortion curve comparing to the existing method.

A method for identifying JPEG 2000 images with different coding parameters, such as DWT-filters, code-block sizes, quantization step sizes, and resolution levels, is presented. The proposed method does not produce false-negative matches regardless of the different coding parameters. Moreover, the proposed method is fast because it uses the number of zero-bit-planes. This number can be extracted from the JPEG 2000 codestream by only parsing the header information. Experimental results showed the effectiveness of the proposed method for digital cinema applications.

In this paper we propose a method of constructing packet-level LDGM codes that offer backward compatibility with conventional viewing devices. Our proposed method makes it possible to watch content even if the viewer does not support any FEC module. Moreover, the method also improves the coding efficiency through the combined use of packet division and interleaving methods. In general, there is a tradeoff between computation complexity and performance, but our proposed method improves coding efficiency by using a message passing decoding scheme that does not require any additional computation. The coding efficiency of the proposed method is evaluated both experimentally and theoretically.

This report proposes an integer color transform for lossless coding of four color components of images. An existing color transform has a fixed set of coefficients and therefore it can't be adaptive to each image. We utilize eigen-vector of the covariance of the four components to increase data compaction performance. We also utilize fixed relation between two green components to simplify computational cost of the transform. It was confirmed that the proposed method reduces entropy rate by 0.4 bit comparing to the existing method.

In this report, we reduce the number of lifting steps of the two dimensional (2D) discrete wavelet transform (DWT) in JPEG 2000 by introducing a factorization of the separable 2D transfer function into some non separable 2D (NS-2D) functions. The order of the 1D lifting steps in vertical and in horizontal are properly replaced so that all of them are synthesized into NS-2D functions. As a result, the number of the lifting steps of the 2D 9/7 DWT is reduced to 75 [%] maintaining compatibility to the JPEG 2000. Since the number of the lifting steps is essentially proportional to a delay of output signals, it contributes to construct a low latency DWT under the assumption that direct accessing to a 2D intermediate memory is permitted.

This report theoretically derives a condition on the word length of coefficients and signals such that the irreversible 9-7 discrete wavelet transform (DWT) in JPEG 2000 becomes lossless for a constant valued (DC) input signal. Under the condition, the output signal contains no error in spite of rounding of coefficients and signals inside the DWT circuit. We have previously analyzed the condition for any value of DC signals. However, it was too strict and the word length under the condition was redundant for a practical case. In this report, we derive a new condition for a specific DC value and show that the minimum word length is halved under the new condition.

We propose double-search-window block matching and describe an FFT-based fast algorithm for it. The double-search-window defines two different-sized search windows, an inside-search-window and an outside-search-window, for block-matching. The proposed method achieves the same accuracy as a direct SSD full search in the inside-search-window. It also offers a pseudo-full-search result in the outside-search-window. By evaluating the results comprehensively, the new approach could achieve more precise motion vectors as well as generate the predicted pictures with fewer matching errors, but with almost the same computational load as in the conventional single-search-window block matching.

This paper proposes a “reversible” three-point Karhunen Loeve transform (KLT) for de-correlation of RGB color components of an image signal. It is composed of “reversible” two point rotation transforms which have singular points (SP) and their rotation angles change depending on correlation of an input signal. When the angle is close to SP, rounding errors inside the transform are magnified to huge amount. To avoid this problem, we introduce permutation “order” and “sign” of signals. We also make it clear that the proposed method can shift SP by 0, 90, 180 or 270 degrees so that distance between SP and the rotation angle is maximized. It was observed that the proposed method reduces variance of the error by 10 [%].

This report proposes a new transcoding between a lossless encoder and a lossy decoder for color image signals. It requires a reversible color transform (Rev.CT) with compatibility to an irreversible color transform (Irrv.CT). An existing Rev.CT guarantees lossless reconstruction of signals. However its compatibility to an Irrv.CT is not high enough for practical use. This is because a scaling is excluded from the Rev.CT for lossless coding. In this report, we embed the scaling into quantization in lossy decoding to realize both lossless coding and transcoding with high compatibility. It is experimentally confirmed that image quality is improved by more than 37.6 [dB]. We also evaluate its performance based on various criteria including, bit-extension, word-length of signals and coefficients.

A new bit allocation scheme for JPEG 2000 systems in parallel-distributed environments is proposed. In recent years, super high-definition images are being used for not only digital cinema also alternative services e.g. real-time sports in theaters. In general, such high-definition images are divided into small segments, such as tiles or components, and compressed. However, this process leads to low coding efficiency. In contrast, our proposed scheme controls the bit allocation ratio for each color-component separately by using the quantizer properties available in the JPEG 2000 standard. Simulation results show that the proposed scheme achieves the same coding performance as the non parallel-distributed approach.

In this report, permutation of order and sign of signals are introduced to avoid singular point problem of a reversible transform. When a transform is implemented in the lifting structure, it can be "reversible" in spite of rounding operations inside the transform. Therefore it has been applied to lossless coding of digital signals. However some coefficient values of the transform have singular points (SP). Around the SP, rounding errors are magnified to huge amount and the coding efficiency is decreased. In this report, we analyze the SP of a three point KLT for RGB color components of an image signal, and introduce permutation of order and sign of signals to avoid the SP problem. It was experimentally confirmed that the proposed method improved PSNR by approximately 15 [dB] comparing to the worst case.

One category of fast full-search block matching algorithms (BMAs) is based on the fast Fourier transformation (FFT). In conventional methods in this category, the macroblock size must be adjusted to the search window size by zero-padding. In these methods, the memory consumption and computational complexity heavily depend on the size difference between the macroblock and the search window. Thus, we propose a novel FFT-based BMA to solve this problem. The proposed method divides the search window into multiple sub search windows to versatilely control the difference between the macroblock and the search window sizes. Simulation results show the effectiveness of the proposed method.

This paper proposes a reversible information hiding for supporting hierarchical control of access to embedded data. The proposed method firstly hides the most important information to an image by the reversible manner in which information is enciphered. To a stego image to which once information is hidden, this method embeds different information that is enciphered with a different key. In the proposed method, only one key corresponding to the most important data is managed, and other keys are generated from the managed key through a one-way hash function. The privileged user receives the managed key and he/she extracts and deciphers all hidden data, whereas the unprivileged user receives a key that for the least important data. The proposed method also has tamper detection ability. Simulation results show the effectiveness of the proposed method.

We propose two fitting functions for shift estimation using discrete cosine transform sign phase correlation (DCT-SPC) with non-integer accuracy. The DCT-SPC can be used in order to estimate the shift values and the similarity between two signals. However, the estimated shift values are limited to integer numbers. The proposed fitting functions enable the DCT-SPC to estimate the non-integer shift values using only the sign of the DCT coefficients. The multiple-peak model in the proposed fitting functions provides more accurate values than other models. Simulations are presented to demonstrate the effectiveness of the proposed fitting functions.

We propose a phase-scrambling method for blind image matching, which is a direct image matching between invisible images. The phase scrambling is motivated by visual protection of images and prevention of illegal image matching. Phase-only correlation (POC) can be directly applied to images protected by the proposed phase scrambling in order to estimate similarity and translation between images. POC with synchronized scrambling provides blind image matching, in which phase scrambling does not affect the accuracy of POC. The effect of visual protection and prevention of illegal image matching is evaluated through simulations to show the effectiveness of the proposed method.

In this report, we propose a reversible two dimensional (2D) discrete wavelet transform (DWT) compatible to the irreversible 9-7 DWT for lossy coding in the JPEG 2000. All the filters and scalings are factorized into lifting steps, and signals are rounded into integers, so that the proposed DWT becomes reversible and applicable to lossless coding. Furthermore, we factorize the separable 2D transfer function of the 2D DWT into non separable 2D functions to reduce the lifting steps and rounding operations. As a result, errors due to the rounding operations are reduced and the compatibility to the irreversible 9-7 DWT is improved.

A bit-stream-level method for identifying encrypted JPEG 2000 images without having to decrypt them is described. It is well known that editing the scenes of the movie is often necessary. In the editing process, image identification plays an important role in finding a frame that should be re-encoded. Thus, identification of encrypted images is very useful in digital cinema because all frames in digital cinema are encoded and encrypted. The proposed method directly uses encrypted JPEG 2000 images so that decryption-free identification without JPEG 2000 decoding is possible. The proposed method is both accurate and fast. In principle, identification based on the proposed method does not produce false negative matches regardless of the compression ratio. Moreover, since there is no need to decode and decrypt the images, the average processing time for identification is very short and independent of the encoded image size.

This paper proposes an efficient access control method for content with multiple dimensions of hierarchical scalability. The proposed method versatilely serves various quality content by simultaneous control in every dimensions of scalability. This method manages single entity for the content and single encipher key, though it prevents users from colluding. A user receives single ciphered content and single decipher key to obtain the content with the permitted quality. In comparison to the conventional collusion attack-resistant access control methods, the sophisticated key generation manner in the proposed method reduces the required length of the managed key and that of the managed codestream.

This paper proposes a new class of lifting wavelet transform which can guarantee losslessness of specific signals, e.g. white balance. The 5/3 wavelet transform composed of two lifting steps can reconstruct an input signal without any loss and has been utilized for lossless coding. The 9/7 wavelet contains two more lifting steps and two scaling pairs for effective lossy coding. However the losslessness is not guaranteed due to rounding of signal values and scaling coefficient values. This paper analyzes condition on word length (WL) and bit depth (BD) for the losslessness and proposes a new class of wavelet transform with "DC lossless " property which is a kind of specific losslessness. This can be utilized as a standard condition for algorithms or LSI processors to guarantee no error from the wavelet transform for white balance signals.

This paper proposes a scrambling method for motion JPEG (MJ) videos and moving objects (MOs) detection from scrambled videos. In the proposed method, both scrambling and MOs detection utilize the property of the positive and negative sign of discrete cosine transformed (DCT) coefficients. Since a DCT sign is encoded separately from its corresponding magnitude, the sign is processed without decoding a MJ video. This feature makes the proposed method fast. Since this method only inverts some of DCT sign bits to scramble videos, the codestream length never change by scrambling. By this scrambling manner, the scrambling strength is easily controlled by the bit inversion ratio. Moreover, this method holds the relation of DCT signs between two frames to be able to detect MOs from scrambled videos. The proposed method can further scramble the detected MOs. In addition, descrambling never require either the position or the shape of the objects to be descrambled.

This paper proposes an encryption method that uses short keys to enable hierarchical access controls for JPEG 2000 codestreams. The proposed method provides images of various quality levels that may be different from the quality at encoding, though it uses a single codestream and a single managed key (masterkey). Only one key generated from the masterkey is delivered to a user authorized to access a reserved quality image. This method also stems users' collusion to access superior-quality images. Some conventional methods of this proposed method serve the above features, but those keys are much longer than the proposed method. The proposed method uses the smaller number of partial keys than the conventional methods.

We propose a method of image registration between scrambled images in order to guarantee the secure data management. This method uses either phase-only correlation or DCT sign phase correlation for the registration. Correlation and registration values between images are directly estimated in the scrambled domain. Information in each image is protected and the relative relation with other images is held by synchronizing scrambling. As a result, the proposed method has no effect on registration accuracy. Translation estimation experiments are performed to show the appropriateness and the effectiveness of the proposed method.

The 5/3 wavelet transform with double lifting steps in JPEG 2000 can reconstruct a signal without any loss. It has been utilized for lossless coding. The 9/7 wavelet transform contains two more lifting steps and scaling operations to improve performance for lossy coding. The loss is due to (1) quantization of band signals, (2) rounding of signals after scaling and (3) finite word length expression of scaling coefficients. This paper analyzes conditions on word length of coefficients and bit depth of rounded signals for no loss. It also proposes a new structure of lifting wavelet by changing order of the lifting step and the scaling. As a result, the rounding error is not scattered by the lifting steps and the error is minimized in mini-max sense.

In this paper, we propose a no-reference (NR) method for estimating the degradation of Motion JPEG2000 (MJP2) video due to packet loss. The proposed method estimates the peak signal-to-noise ratio (PSNR) between received MJP2 frames affected by packet loss and the originally encoded frames without having information about the latter. The PSNR of a frame could be estimated in the wavelet domain, namely before the JPEG2000 decoding process, by utilizing adjacent frames that possess similar amounts of wavelet energy. Simulation results showed that the estimated PSNR is very close to the true value obtained by the full-reference calculation. Our method can be used to monitor the video quality in an IP-based transmission of MJP2, as well as for automatic judgment of error control in streaming applications, or in video broadcasting. For example, if the PSNR of a frame does not reach a predefined level, the frame will be rejected before the decoding process or it will be processed by a concealment techniques before being displayed.

A close relationship between the sign information of discrete cosine transform (DCT) coefficients and the phase term of discrete Fourier transform coefficient is explained. From this relationship, the DCT sign-only correlation is proposed and its relationship with the phase-only correlation (POC) is considered. As a result, the proposed correlation can be applied to applications based on the POC, and also provide a theoretical explanation of target image search and retrieval using the sign information of DCT.

We propose a fast method to retrieve images from a JPEG image database. The proposed method is intended to retrieve similar images, including their compressed versions, and to identify exact match and all compressed versions of a query image simultaneously. Similarity level is measured based on the non-zero DCT coefficients signs, which serve as features. The method is simple and fast because the DCT coefficients signs can be obtained by only entropy-decoding the bitstream. There is no need to calculate features explicitly. Furthermore, the method is robust to JPEG compression, particularly for image identification purpose.

This paper proposes a reversible data hiding method that embeds an L-level data sequence to images in the spatial domain. Though reversible data hiding once distorts the image to hide data into it, the distorted image is completely separated to the original image and the hidden data. The proposed method uses only one parameter to embed and extract data, and it extracts data without any location map. In addition, it can control embedding capacity according to payload for suppression of embedding distortion. Simulation results show the effectiveness of the proposed method.

This paper proposes a collusion attack-resilient method of encryption for access control of JPEG 2000 codestreams with hierarchical scalabilities. The proposed method generates one encryption key from one single key by multi-dimensional scanning to serve encryption keeping the scalability of codestrems. To avoid collusion attacks in which multiple users generate an illegal key from their own keys to overcome the access control, sufficient conditions are considered in this method. Moreover, a skip encryption is introduced to decrease the computational complexity and key management-and-de-livery cost of encryption. Simulation results show the effectiveness of the proposed method.

This paper proposes a tamper detecting method for images using mathematical morphology. The proposed method utilizes the idempotent property of morphological operations rather than fragile watermarking methods. While fragile watermarking methods that must extract the embedded watermark and compare it with the possible watermark to detect tampers, the proposed method detects and localizes tampers by only morphological operations and image subtraction in the spatial domain. Moreover, a visual signature system is proposed based on the proposed tamper detecting method. Experimental results show the effectiveness of the proposed method.

This paper proposes a moving objects scrambling method for Motion JPEG movies that consists of two parts: a moving objects detection based on DCT sign-only correlation (DSOC) and a partial scrambling. DSOC is the correlation of the positive and negative signs of DCT coefficients. Utilizing the codestream structure of Motion JPEG movies and the relation between motion and DSOC, the proposed moving objects detection is achieved in codestream domain. The proposed scrambling is also achieved in codestream domain by inverting signs of DCT coefficients that each sign is independently encoded as one bit. The codestream domain processing in this method serves low processing time and keeps coding efficiency. Moreover, the proposed method completely descrambles scrambled codestreams to the original without any knowledge on the position and shape of moving objects. Simulation results show the effectiveness of the proposed method.

We propose two identification methods for JPEG-coded images. The purposes are to identify the images that are compressed from the same original image with various compression ratios in fast and robust manner. The first approach can avoid identification leakage or false negative (FN), and could result in a few false positives (FP). The second approach can avoid both FN and FP, with a slightly longer processing time. By combining the two schemes, a faster and a more perfect identification can be achieved, in which FN and FP can be avoided

An encryption scheme is proposed that considers hierarchies in media, such as text, images, sound, and so on, in a composite multimedia content to enable versatile access control. In the proposed scheme, a content provider has only one managed key (the master key) for a particular multimedia content, and an user who is permitted to access a reserved content objects in the composite content receives only one key that is subordinately generated from the master key. The key generated from the identical master key is delivered to another user, and this permits the user to access different objects. This scheme introduces a new key concept, namely "unusable key", to keep all objects encrypted in a particular medium and to simultaneously decrypt several objects in other media. The other new key, "numbering key," is also used in this scheme to support simultaneous partial decryption of multiple images that are coded with scalable coding technology. Simulation results show the effectiveness of the proposed scheme.

This paper proposes a watermarking method that hides an integer in an image by modifying the sum of the quantized discrete cosine transformed coefficients. Because modulo arithmetic-based modulation is applied to an integer to be embedded, less image distortion is achieved. Moreover, coefficients not being chosen to embed an integer are restored to their pre-quantized form, even though they are once quantized. The image-quality of a watermarked image is theoretically analyzed and is found to be superior to those of image-quality guaranteed watermarking methods.

This paper proposes an encryption scheme maintaining hierarchy of JPEG 2000 codestreams for flexible access control. JPEG 2000 generates hierarchical codestreams and has flexible scalability, such as, SNR, space, or color components. In a proposed method, only one managed key (master key) exists, and for a user permitted to access a reserved image quality, only one key generated by the master key is delivered. Also, a different key generated by the master key is delivered for a user permitted to access a different quality. The proposed access control method is available for all kinds of JPEG 2000 scalability. On the other hand, some conventional methods need several keys to control image quality. Another conventional method is not suitable for JPEG 2000 scalability.

This paper proposes a quantization-based watermarking method that generates watermarked images of a desired image-quality independently of the characteristics of the original images. To guarantee the image-quality, this method utilizes the relationship between a transformed domain and the spatial domain for the signal energy. The proposed method extracts an embedded watermark sequence from a watermarked image without referring to any original images.

This paper proposes an efficient packet loss protection strategy for transmission of JPEG2000 coded images over a packet erasure channel that backwardly compatible with the JPEG2000 part-1 decoder. The proposed scheme utilizes FEC at packet-level, the JPEG2000 layered structure and its error resilient functionalities. Prior to transmission, packets containing initial layers of the code-stream, which is associated lo a minimum acceptable image quality, are protected using a packet-level FEC. With a slight modification in the SOT marker segment and careful placement of the redundant packets, the generated codestream that carried redundant data is compliant with a standard decoder.

A new method for partial-scrambling of JPEG 2000 images based on public-key encryption is proposed. By using public-key encryption, the proposed method provides an easier way of managing the encryption key compared with the secret-key based method and also provides tamper resistance against attacks. Although public-key encryption is usually very time-consuming, the proposed method achieves fast encryption by controlling the number of bytes to be encrypted. An encrypted JPEG 2000 image generated by the proposed method has backward compatibility with a standard JPEG 2000 image, so that it can be decoded using a standard JPEG 2000 decoder. The proposed method also has scalability as to the degree of scrambling on the basis of JPEG 2000 coding units, i.e.. layers. DWT-levels, subbands. or code-blocks.

A method is described for efficient partial-scrambling of JPEG2000 images that avoids generating marker codes and improves the ability to control the degree, strength, and computational complexity of scrambling. This higher control ability is due to the use of a parameter. This parameter also controls the scrambling time, an important consideration for real-time processing.

A method of error correction for JPEG2000 images is proposed in this paper. The method uses the layer structure that is a feature of the JPEG2000 and an error correction code. The upper layers of the code stream are coded using an error correcting code, and the parity data are hidden in the lowest layer. The hidden data are used for error correction at the decoder. Several error correction codes with different strength are selected for the main header, packet headers, and bodies. Since the resulting code stream has the same data structure as a standard JPEG2000 code stream, it can be decoded with a general decoder. Simulation results demonstrated the effectiveness of the proposed method.

This paper proposes a fast implementation technique for RLS adaptive filters. The technique has an adjustable parameter to trade the throughput and the rate of convergence of the filter according to the applications. The conventional methods for improving the throughput do not have this kind of adjustability so that the proposed technique will expand the area of applications for the RLS algorithm. We show that the improvement of the throughput can be easily achieved by rearranging the formula of the RLS algorithm and that there are no need for faster PEs for the improvement.

We proposed a method for embedding binary data into JPEG bitstreams and extracting embedded data from JPEG bitstreams using the standard JPEG decoder. In the proposed method, we can decode the image from the embedded binary data JPEG bitstreams first using the traditional standard JPEG decoder, and then we can extract the embedded binary data perfectly by the post-processing from the decoded JPEG image. For the post-processing, we use only the decoded image data to extract the embedded binary data. Namely, we do not need any kind of particular parameters, which are used for JPEG decoding, such as quantization table value. Thus, we can use the traditional standard JPEG decoder for the pre-processing of extracting binary data. Furthermore, we address the effect of the calculation bit accuracy of the DCT and inverse discrete cosine transform (IDCT) for extracting embedded binary data perfectly as post-processing. Simulations using extracting embedded binary data as post-processing are presented to quantify some performance factors of concern. We confirmed that the proposed method could be of practical use.

In the MPEG standard, motion estimation (ME) is used to eliminate the temporal redundancy of video frames. This ME is the most time-consuming task in the encoding of video sequences and is also the one using the most power. Using low-bit images can save the power of ME and a conventional architecture fixed to a certain bit width is used for low-bit motion estimation. It is known that there is a trade-off between power and image quality. ME may be used in various situations, and the relation between demands for power or image quality will depend on those circumstances. We therefore develop an architecture for a low-bit motion estimator with adjustable power consumption. In this architecture, we can select the bit width for the input image and adjust the amount of power for ME. To evaluate its effectiveness, we designed the motion estimator by VHDL and used the synthesis results to estimate the performance.

We propose a method for inserting optional binary data into MPEG bitstreams. By inserting additional information like index label effectively into video database, we can deal the additional information with video data itself easily. The proposed method allows us to extract the inserted data perfectly though the MPEG video is a lossy algorithm, and the quality of the decoded images after extracting additional information is almost the same as that of the ordinary MPEG bitstreams. Moreover, the standard MPEG decoder can be used for decoding the bitstreams with inserted binary data. In that case, the additional binary data remains in bitstreams but can decode images without obvious image degradation. To verify the effectiveness of our method, we show simulation results for binary data insertion.

In MPEG-4, motion vectors on both 16/spl times/16 macroblocks and 8/spl times/8 blocks could be used. We propose a block matching algorithm which can find both of them with small computational complexity. It is shown that the number of operations for motion estimation (ME) can be reduced by an expanded successive elimination algorithm (SEA) while maintaining the same accuracy of the full search ME algorithm. In the SEA, the efficiency of reducing the number of ME operations mostly depends upon the appropriate selection of an initial motion vector for searching. We first propose the method of the initial motion vector selection for SEA. We then expand the traditional SEA for the MPEG-1 or MPEG-2 for MPEG-4 in which both 16/spl times/16 macroblocks and 8/spl times/8 blocks are used. In consequence of our simulation, the significance of the proposed method is verified.

This paper proposes an adaptive algorithm, which can be pipelined, as an extension of the delayed least mean square (DLMS) adaptive algorithm. The proposed algorithm provides a capability to achieve high throughput with less degradation of the convergence characteristic than the DLMS algorithm. An architecture for pipelined implementation of the proposed algorithm is considered, and based on this, the conditions for the implementation are derived. An efficient implementation of the architecture with less hardware is also considered.

We propose a method for achieving both a fast convergence speed and the low orders of adaptive digital filters (ADFs) for subband ADFs. The proposed method is based on the multirate repeating method which uses wasted signals by downsampling. First we show how to apply the multirate repeating method to the standard structure of subband ADFs. Next we consider a new structure of subband ADFs for extending the multirate repeating method. Finally, we show the validity of the proposed method by computer simulations.

In this paper, we propose two multidimensional parallel processing methods for rational sampling lattice alteration. Our proposed methods enable us both to implement the rational lattice alteration with the parallel processor approach and to eliminate the redundancy caused by up- and down-sampling. Those methods are provided by extending two conventional block processing techniques for FIR filtering: the overlap-add method and the overlap-save method, respectively.

The proposed algorithm updates filter coefficients using the time averaged gradient vectors and it ensures the good performance under noisy conditions and under finite precision implementation. It is shown that we can update the filter coefficients as many as the algorithm requires to converge even if only short data are available so that the proposed algorithm can converge faster than other gradient type algorithms, i.e. the LMS algorithm. Moreover, LMS algorithm can be regarded as a special case of the proposed algorithm and the tracking ability of the algorithm is shown.>

In this paper, we propose an equivalent model for subband adaptive systems, in order to analyze the performance of the system. Using the proposed model, we obtain a new interpretation of the system, and show theoretical equations of the optimum adaptive digital filter (ADF) coefficients and the least mean squared error (LMSE). From these equations, we also derive the theoretical values of the optimum ADF coefficients and the LMSE. Comparing them with results of computer simulations without the model, we can see agreements of the theoretical values and the simulation results.>

The overlap-add method (OLA) and overlap-save method (OLS) are well known as efficient schemes for high-order FIR filtering. In this paper, new sampling rate conversion methods are proposed by extending the OLA and OLS, and eliminating the redundancy caused by the conversion. First, for finite-duration sequences, a rate conversion with the DFT-domain approach is discussed. Next, using the result, the extended OLA and OLS are proposed for infinite-duration sequences. Finally, the computation complexities of our proposed methods are shown.>

This paper describes a technique for improving the convergence speed of the LMS algorithm. The main idea of the technique is to increase the number of updatings of the filter weights per unit time so that the convergence speed is accelerated according to the number of updatings. The validity of the technique is shown by considering the deriving process of the LMS algorithm from the steepest descent algorithm. We present the two example structures for implementing the technique in actual applications. Finally we show the results of computer simulations.>

Conventional subband adaptive filters (ADF) using filter banks have shown degradation in performance because of the nonideal nature of filter banks. For this problem, the authors propose an alias free subband structure for adaptive filtering using polyphase frequency sampling filter (FSF) banks. As a preliminary, they make it clear that the conventional polyphase discrete Fourier transform (DFT) bank is equivalent to a class of the FSF bank. Then, they propose a new class of ADF using the FSF banks based on the frequency sampling theorem. As a result, the proposed technique enables subband adaptive filtering without the degradation effect of both the aliasing and cross terms, even if one chooses critical subsampling.>

Describes the development of a technique for subband image coding called the symmetric extension method which utilizes the nature of a symmetrically extended image to achieve high quality coding. There are, however, some restrictions on the usable subband coding system. The development is done based on the property of a symmetrically extended signal in the analysis process to remove the restriction. The behavior of the sequence after two processes, filtering and decimation, is examined and formulated. The development extends the application area of the technique.

Some researchers have explored the use of filter banks to improve the performance of adaptive algorithms. However, some of these approaches have shown degradation in performance. A new class of filter bank (ideal orthogonal filter bank) is defined, which enables one to avoid the degradation, even if one chooses critical subsampling in adaptive filtering. A filter bank (pseudo OFB) whose frequency response is the same as that of the ideal OFB at discrete frequency points is proposed. As a result, using the proposed filter bank, one can identify an unknown system at the discrete frequency points.>

One of the problems with subband image coding is the increase in image sizes caused by filtering. To solve this, it has been proposed to process the filtering by transforming an input sequence into a periodic one. Then, filtering is implemented by circular convolution. Although this technique solves the problem, there are very strong restrictions, i.e. limitations on the filter type and on the filter bank structure. The development of this technique is presented. Consequently, any type of linear-phase finite impulse response (FIR) filter and any structure of filter bank can be used.>