Supervised learning of deep neural networks heavily relies on large-scale datasets annotated by high-quality labels. In contrast, mislabeled samples can significantly degrade the generalization of models and result in memorizing samples, further learning erroneous associations of data contents to incorrect annotations. To this end, this paper proposes an efficient approach to tackle noisy labels by learning robust feature representation based on unsupervised augmentation restoration and cluster regularization. In addition, progressive self-bootstrapping is introduced to minimize the negative impact of supervision from noisy labels. Our proposed design is generic and flexible in applying to existing classification architectures with minimal overheads. Experimental results show that our proposed method can efficiently and effectively enhance model robustness under severely noisy labels.

Speckle noise is an inherent disturbance in coherent imaging systems such as digital holography, synthetic aperture radar, optical coherence tomography, or ultrasound systems. These systems usually produce only single observation per view angle of the same interest object, imposing the difficulty to leverage the statistic among observations. We propose a novel image restoration algorithm that can perform speckle noise removal without clean data and does not require multiple noisy observations in the same view angle. Our proposed method can also be applied to the situation without knowing the noise distribution as prior. We demonstrate our method is especially well-suited for spectral images by first validating on the synthetic dataset, and also applied on real-world digital holography samples. The results are superior in both quantitative measurement and visual inspection compared to several widely applied baselines. Our method even shows promising results across different speckle noise strengths, without the clean data needed.

A novel quality of service (QoS) provisioning algorithm for home networks is presented in this paper. The algorithm carries out the QoS-aware bandwidth allocation using the general regression neural networks (GRNNs). Among all the allocations predicted to receive positive service responses, the algorithm finds the allocation with minimum total bandwidth for the current service. The service response prediction is based on the GRNN with the training set containing the bandwidth allocations and their service responses for past transmissions. The new service responses will then be used to update the training set for the subsequent transmissions. To attain accurate tracking of diversified service requirements, flexible specification of service response levels and QoS levels are provided. Both analytical and numerical studies reveal that the proposed algorithm is able to provide prompt or steady reactions to the service feedback depending on the variations of the source data rates. Because of its simplicity and effectiveness, the proposed algorithm is well suited for dynamic QoS management for heterogeneous home networks.

This paper introduces a novel FPGA-based signature match co-processor architecture serving as the core of a hardware-based network intrusion detection system (NIDS). The signature match co-processor architecture is based on the shift-or algorithm. The architecture is comprised of simple shift registers, or-gates, and ROMs where patterns are stored. As compared with related work, experimental results show that the proposed work achieves higher throughput and less hardware resource in the FPGA implementations of NIDS systems

With the appearance of multi-gigabit network infrastructure, a typical network intrusion detection system (NIDS) has to cope with the network speed. By examining each packet flowing through a network segment, suspicious packets are detected and reported to assure security. Up to 57% of the execution time in a NIDS is found to compare string against a predefined/known pattern. It is hard to implement a multi-gigabit performance NIDS without hardware support. This paper proposes a very high speed string matching algorithm which can be easily implemented into FPGAs. The parallel matching design takes a segment of text from the payload of a packet and detects all possible tokens including those crossing text segment boundaries. Simulation results show a throughput of 23.43 Gbps with a moderate operating frequency of 366.2 MHz

This paper presents a new video-coding technique based on displaced frame difference (DFD) prediction and multiresolution block matching for layered transmission. Starting from the base layer, the algorithm encodes one layer at a time until the design of the top layer is completed. For the encoding of each layer, the DFD prediction and multistage motion-vector quantization (MSMVQ) techniques are employed so that the results of DFD encoding and block matching at previous layers can be used to enhance the coding efficiency at that layer. Numerical results show that the algorithm is effective for the applications where high rate-distortion performance and low computational complexity for layered video transmission are required.

In this paper, a novel medical data compression algorithm, termed layered set partitioning in hierarchical trees (LSPIHT) algorithm, is presented for telemedicine applications. In the LSPIHT, the encoded bit streams are divided into a number of layers for transmission and reconstruction. Starting from the base layer, by accumulating bit streams up to different enhancement layers, we can reconstruct medical data with various signal-to-noise ratios (SNRs) and/or resolutions. Receivers with distinct specifications can then share the same source encoder to reduce the complexity of telecommunication networks for telemedicine applications. Numerical results show that, besides having low network complexity, the LSPIHT attains better rate-distortion performance as compared with other algorithms for encoding medical data.

Transmission of compressed images is challenging because the communication channels can be noisy and the compressed bitstreams are sensitive to channel errors. Most of the proposed methods in the literature do not guarantee the quality of the received image over noisy channels. We propose an ARQ-based diversity system as a solution, in which the use of the diversity system avoids the problems when one or some channels get congested, and the ARQ scheme guarantees the quality of the received image. A specific ARQ-based diversity system is designed for transmission of images compressed using EZW (embedded zerotree wavelet). Experiments performance results are shown.

We develop a multi-channel channel-optimized scheme for embedded zerotree wavelet (EZW) image compression in a noisy transmission environment. A block-based modification for EZW is applied to improve the robustness of EZW, and to produce several coded bitstreams for transmission over multiple channels with different noise conditions. Then the respective channel noise is considered in the rate-distortion analysis, and the resultant rate-distortion functions are used for optimal bit allocation among the coded bitstreams. The case of burst noise is analyzed as an example.

We apply the dynamic bit allocation to the block-based EZW algorithm for robust image compression. To optimize the performance of the bit allocation, the effect of the channel noise is concerned. The robustness of our method was evaluated.

A supervisory fuzzy neural network (FNN) control system is designed to track periodic reference inputs in this study. The control system is composed of a permanent magnet (PM) synchronous servo motor drive with a supervisory FNN position controller. The supervisory FNN controller comprises a supervisory controller, which is designed to stabilize the system states around a defined bound region and an FNN sliding-mode controller, which combines the advantages of the sliding-mode control with robust characteristics and the FNN with online learning ability. The theoretical and stability analyses of the supervisory FNN controller are discussed in detail. Simulation and experimental results show that the proposed control system is robust with regard to plant parameter variations and external load disturbance. Moreover, the advantages of the proposed control system are indicated in comparison with the sliding-mode control system.

A new fast codeword search algorithm for vector quantizers (VQ) is presented in this paper. This algorithm uses the pyramid structure of the codewords to accelerate the encoding process. The pyramid structure is obtained using the wavelet transform. This algorithm is able to reduce the codeword search time without sacrificing the performance and storage complexity of the VQ. Simulation results show that this algorithm is well-suited for the VQ with high vector dimension.

This paper presents a new progressive image transmission (PIT) design algorithm in which the resolution and resources (rate or distortion and storage size) at each transmission stage are allowed to be prespecified. This algorithm uses the wavelet transform and tree-structured vector quantizer (TSVQ) techniques. The wavelet transform is used to obtain a pyramid structure representation of an image. The vector quantizer technique is used to design a TSVQ for each subimage so that all the subimages that constitute the image at the current stage can be successively refined according to the resources available at that stage. The resources assigned to each subimage for the successive refinement is determined to optimize the performance at the current stage under the resource constraints. Normally, the resource constraints at each stage are determined by the specification of the transmission time or distortion for image data and the storage complexity of the TSVQ. The resolution at each stage is determined/specified according to the application or as part of the design process to optimize the visual effect.