The paper presents a brief review of the current standardization activities in Video Coding for Machines (VCM), including the MPEG 2022 Call for Proposals for VCM. These activities and the conditions defined for this call are the background of the proposed technology for VCM. In this technology, a search for Regions of Interest (RoIs) is implemented in the encoder in order to find the areas where the objects are possibly located. The background (the areas outside RoIs) is simplified and retargeting is executed before the video is compressed. For low bitrates, the experimental results demonstrate significant improvements of the results for object detection, instant segmentation and object tracking as compared to the state-of-the-art VVC coding.

A typical learning-based video compression scheme consists of motion coding and residual coding. In this paper, our deep video compression features a motion predictor and refinement networks for interframe coding. To save the bits for transmitting motion information, our scheme performs local motion prediction and sends only the differential motion vectors to the decoder. In the residual coding, we couple the residual decoder with the refine-net to reduce residual signal bits. The experiments show that our work can produce a very competitive coding performance compared to the other learning-based predictive video codecs.

The paper deals with AVC, HEVC and VVC video encoder modeling using a universal formula for the output bitrate as a function of the quantization step. It is shown that this model is relatively accurate and may be used for bitrate control. Moreover, from the model for AVC, the models for HEVC and VVC can be quickly derived, thus reducing the processing time needed for VVC model parameters derivation. The model is applicable to multiview plus depth video coding, both using simulcast and MV-HEVC or 3-D HEVC. The applications to bitrate control for coding of multiview plus depth video are considered together with the relevant experimental data.

Virtual reality, free-viewpoint television, virtual navigation, 360° video are the areas of research and technology that need efficient compression of multiview video plus depth acquired by cameras with arbitrary positions. Astonishingly, proliferation of 3D extensions of AVC and HEVC technology is very low. Therefore, in this paper, we present a study on independent coding of views and depth maps. A simple technique is proposed to estimate quantization step for depth as a function of the quantization step for multiview video. This technique is studied in the context of multiview video plus depth acquired using cameras located around a scene. The approach is based on simple modeling of the relation between quantization parameters for depth and multiview video. The experimental results are obtained for stereoscopic video with two respective depth maps. For standard MPEG test sequences, the results demonstrate usefulness of the approach for HEVC, VVC, MV-HEVC codecs.

The paper reports a new view synthesis method for omnidirectional video with the ability to slightly displace a virtual viewpoint, i.e. the paper describes a novel synthesis method for 3DoF+ 360 video. This new method is noteworthy because of its high versatility and reliability: the method is appropriate for both perspective and omnidirectional input views, is able to render both perspective and omnidirectional views, and the produced synthetic views differ from the respective ground truth images less than with other view synthesis methods. These important features result from several innovations, e.g., prioritization of input views, efficient inpainting adapted to equirectangular projections, and efficient color correction. The experimental results demonstrate that the new method outperforms the state-of-the-art methods used hitherto.

Delays in the delivery of immersive video to Head Mounted Devices (HMDs) are considered in the paper. The goal of this paper is to analyze the impact of video streaming latency on user experience with HMDs. The paper reports the results of the subjective quality assessment as a function of delay. The possible practical solutions are identified and, on this base, an experimental model of the considered systems is proposed. In order to properly test the influence of the streaming delay on the quality, the extensive subjective tests have been performed. Interesting conclusions have been drawn, showing that human acceptance of delay of translation of virtual point of viewing is much stronger than for delay of rotation of virtual viewing direction in immersive video. Finally, the influence of such observations on the system architectures is concluded.

In this paper, we identify the general requirements for network-on-chips (NoCs) and the general characteristics of field-programmable gate arrays (FPGAs) from leading producers. Based on the analysis provided, an FPGA-oriented NoC called RingNet is proposed. As a distinctive feature, RingNet uses communication through a centrally placed memory that aims at preventing network congestions and limiting the network buffer requirements. Optimal utilization of FPGA resources is one of the goals of RingNet development. Especially buffers are implemented in distributed RAM available in FPGAs, and the virtual cut-through is used as an efficient switching technique for FPGA. Simulations prove guaranteed throughput, predictable latency, and fair network access provided by RingNet. Synthesis results for sample FPGAs from Xilinx, Intel, and Lattice prove the universality of RingNet. The provided analysis of NoC implementations leads to the conclusion that RingNet needs fewer resources and supports higher clock frequencies than the widely used AXI4 architecture.

In this paper we present a novel method for depth estimation from stereoscopic 360-degree video. As for now, no such methods have been presented in the literature. First we show mathematical foundations for correspondence matching polar coordinates. Then, the attained formulas are simplified to show relationship to depth estimation in rectified pairs of rectangular images. The proposed method has been implemented in Depth Estimation Reference Software (DERS) developed by Motion Picture Experts Group (MPEG) of International Organization for Standardization (ISO) and the International Electrotechnical Commission (IEC). Finally, we present experimental results in form of depth maps estimated for “Dancing 360” stereoscopic omnidirectional sequence.

Free-viewpoint television (FTV) and virtual navigation appear to be hot research topics. In this paper, the authors study the practical development of free-viewpoint television systems that provide the functionality of virtual horizontal navigation around real scenes. The considerations are focused on practical systems that use purely optical depth estimation and might be employed in the next few years. The architectures of such systems are discussed in detail, including acquisition, preprocessing, depth estimation, compression, and presentation. In particular, the optimization of camera locations is discussed, and it is shown that video acquisition using camera pairs is advantageous for scenes with a substantial amount of occlusions. The theoretical considerations are supported by experimental results obtained for standard test multiview video sequences. Furthermore, the paper describes FTV video acquisition systems that consist of modules with pairs of cameras. The modules are sparsely located in arbitrary positions around a scene. Each camera module is equivalent to a video camera with a depth sensor. The hardware requirements, video processing algorithms, and experimental results are reported. In particular, for such systems, a compression technique is discussed that is more efficient than the new three-dimensional HEVC technology. The paper also describes new test video sequences that are obtained from the camera pairs sparsely distributed around scenes.

The demonstration is aimed at presenting a simple free-viewpoint television system that is under development at Poznań University of Technology, Poznań, Poland. Video acquisition using sparsely distributed pairs of video cameras, the depth estimation from video, the virtual video rendering on a server feature the perspective FTV system. The original depth estimation and view synthesis algorithms ensure good quality of the virtual views generated during virtual walks around a scene. The system is developed for broadcasting sport (e.g. judo, karate, volleyball) and cultural (e.g. amateur or professional theater performances) events, as well as for interactive courses and manuals (medical, cosmetics, dancing, technical etc.).

This paper presents a new depth estimation method for multiview systems with arbitrary camera locations. The method exploits the graph cuts method, where vertices of the graph represent segments used for controlling the trade-off between the quality of depth maps and the time of estimation, while preserving the original resolution of a depth map. Moreover, the inter-view consistency of the depth maps, crucial for free-viewpoint television systems, is ensured by introduction of suitable connections in the optimized graph. It makes the proposed method the first that allows generation of spatially-consistent multiview depth maps using segmentation-based estimation. A new method of the adaptive calculation of the smoothing coefficient was also presented. The performance of the proposed algorithm was tested and compared with the state-of-the-art DERS method, showing an significant improvement, both in terms of the depth maps fidelity and the time of estimation.

The paper presents the concept of a practical free-viewpoint television system with purely optical depth estimation. The system consists of camera modules that contain pairs or triples of cameras together with the respective microphones. The camera modules can be sparsely located in arbitrary positions around a scene. Each camera module is equivalent to a video camera with a depth sensor and microphones. The hardware requirements, the video and audio processing algorithms and the preliminary experimental results are reported. In particular, for such systems, a compression technique is discussed that is more efficient than the new 3D-HEVC technology. A set of new test sequences obtained with the use of camera pairs are presented.

In the paper, we propose a new method for the virtual view synthesis called as Multiview Synthesis. In our approach, virtual views are synthesized using two neighboring real views, but the disoccluded areas are not inpainted, but filled by the information from the further real views - additional views, traditionally not included in the view synthesis. The whole synthesis is performed with triangles rather than with individual pixels. In the proposed Multiview Synthesis, additional steps of adaptive color correction, blurred edge removal and spatial edge blurring are also included. The experimental results show that both the objective and subjective quality of the synthesized views is significantly higher than the quality of views obtained using the state-of-the-art MPEG reference view synthesis software.

We deal with the processing of multiview video acquired by the use of practical thus relatively simple acquisition systems that have a limited number of cameras located around a scene on independent tripods. The real-camera locations are nearly arbitrary as it would be required in the real-world Free-Viewpoint Television systems. The appropriate test video sequences are also reported. We describe a family of original extensions and adaptations of the multiview video processing algorithms adapted to arbitrary camera positions around a scene. The techniques constitute the video processing chain for Free-Viewpoint Television as they are aimed at estimating the parameters of such a multi-camera system, video correction, depth estimation and virtual view synthesis. Moreover, we demonstrate the need for new compression technology capable of efficient compression of sparse convergent views. The experimental results for processing the proposed test sequences are reported.

Rate-Distortion Optimized Quantization (RDOQ) is an encoding optimization technique that may be modified with no implications to the bitstream compliance with the standard. In the paper, we propose an RDOQ variant with exact cost estimation resulting in higher compression ratios as compared to HEVC Test Model. This improvement is obtained at the cost of the complexity increase that may be scaled by replacing the individual steps of exact estimation by the simplified ones from HEVC Test Model. In that way, a family of RDOQ variants is created. The performance of the exact variant from this paper defines a kind of the performance limit for RDOQ.

In the paper, novel methods for quality control in multiview with depth video compression are described. First, a simple steepest-descent-based algorithm for adjusting quantization parameters for texture and corresponding depth to obtain the maximal quality for a given bitrate is proposed. Then, experimental results are shown, basing on which a mathematical model for calculation of quantization parameters for the depth and texture is proposed. The paper also covers subject of coding depth with different resolution than the texture. The work is presented in context of the most contemporary compression methods related to AVC, but could also be easily adapted to upcoming 3D-video coding extensions of HEVC.

Recently, stereoscopic video systems are studied in the context of many new applications. This paper addresses the question how the road traffic surveillance may benefit from stereoscopic video analysis. In particular, we propose a novel method for estimating vehicle dimensions using stereoscopic video analysis. This method is based on active appearance model (AAM), a technique already used, e.g. for face detection. Experimental results are provided in the paper in order to demonstrate satisfactory precision of the measurements.

The paper deals with nonlinear transformations of depth-sample values and their applications to multiview-plus-depth video coding. It is proposed to apply such a transformation prior to depth coding. For this transformation, the experiments demonstrate improved quality of the virtual views that are synthesized with use of the decoded depth. Such experiments were made for two versions of the prospective extensions of the Advanced Video Coding (MPEG-4 part 10, ITU-T Rec. H.264). standard. The experimental results are given for the scenario where multiview video is encoded independently from depth and also for the scenario where multiview video is coded jointly with depth.

A computationally inexpensive watermark schema for MPEG-4 AVC/H.264 streams that use local mark strength adjustment minimizing visible artifacts is presented. The local mark strength adjustment mechanism and watermark embedding mechanism operate on residual data. This approach minimize the number of operations needed for video bitstream decoding and subsequent watermarked video data encoding. The experimental results included herein, described for real-world MPEG4 AVC/H.264 encoded HDTV video bitstreams prove its robustness to complex attacks, especially camcorder recording attack.

We propose a new coding technology for 3D video represented by multiple views and the respective depth maps. The proposed technology is demonstrated as an extension of the recently developed high efficiency video coding (HEVC). One base views are compressed into a standard bitstream (like in HEVC). The remaining views and the depth maps are compressed using new coding tools that mostly rely on view synthesis. In the decoder, those views and the depth maps are derived via synthesis in the 3D space from the decoded baseview and from data corresponding to small disoccluded regions. The shapes and locations of those disoccluded regions can be derived by the decoder without any side information transmitted. To achieve high compression efficiency, we propose several new tools such as depth-based motion prediction, joint high frequency layer coding, consistent depth representation, and nonlinear depth representation. The experiments show high compression efficiency of the proposed technology. The bitrate needed for transmission of two side views with depth maps is mostly less than 50% than that of the bitrate for a single-view video.

In this paper, we present the most efficient coding tools that are used in a new video compression technology for multiple views with the depth maps. This very well performing technology was designed and developed in response to MPEG Call for Proposals on 3D Video Coding Technology. The proposed technology exploits a tool that reduces the side views and the side depth maps to small disoccluded regions. This way only one central view and one depth map are coded in the HEVC syntax while the remaining views and the depth maps are synthesized in the decoder from the small disoccluded regions and from the central-view data. Therefore, the bitrate needed for a side view is mostly below 20% of the bitrate for single-view video.

The paper deals with multiview video coding using the new technology of High Efficiency Video Coding (HEVC). Implementation of multiview video coding in the framework of HEVC is described together with new specific tools proposed by the authors. Extensive experimental results are reported for compression performance comparison of MVC (ISO 14496-10), HEVC simulcast and two versions of proposed “multiview HEVC”. For “multiview HEVC” the results indicate significant bitrate reduction of about 50%, as compared to the state-of-the-art MVC technology standardized as a part of AVC (MPEG-4, H.264).

During the last two decades, a new technology generation of video compression was introduced about each 9 years. Each new compression-technology generation provides halving of necessary bitrates as compared to the last previous generation. This increasing single-view compression performance is related to increasing compression performance of multiview video coding. For multiview video with associated depth maps, additional significant bitrate reduction may be achieved. The paper reports the original compression technology that was designed and developed at Poznań University of Technology in response to MPEG Call for Proposals on 3D Video Coding Technology. The main idea of this technique is to predict very efficiently the side views and the depth maps from the base view.

The paper deals with efficient coding of motion data for compression of multiview video with depth maps. This research was done in the context of the new compression technology called High Efficiency Video Coding. In the proposed approach, motion vectors and reference frame indices are predicted from reference view for each pixel using the Depth-Based Motion Prediction (DBMP). The new codec with DBMP compression tool was compared to the multiview HEVC-based codec i.e. a set of HEVC codecs augmented with inter-view texture prediction tools adopted from classic multiview codec (MVC) based on the AVC technology. Extensive experiments show the potential of the new DBMP predictor to reduce bitrate up to 12% against the multiview HEVC-based codec.

This paper describes a technique for inter-view depth map consistency improvement for automatically and semi-automatically estimated depth maps. The goal is to improve 3D scene representation consistency by exchanging spatial information between all depth maps in a multiview sequence. Presented technique is based on iterative inter-view information exchange followed by depth quality assessment stage which prevents depth quality loss. The depth-map consistency improvement yields in better multi-view compression ratio and virtual view quality.

The paper describes watermarking of compressed video bitstreams. This technique is proven to be very fast and robust against various attacks including camcorder recording attack. The watermark is embedded using indirect modifications of transform coefficients in selected P- and B-frames. In the watermarked bitstream there exist sequences of watermarked pictures interleaved with sequences of unmodified pictures. The unmodified pictures serve as a reference for the watermark detection. The experimental results described for real-world MPEG-4 AVC/H.264 encoded HDTV video bitstreams are included.

This paper presents a novel compression tool that improves compression performance of Multiview Video Coding when depth information is available for the reference view. This tool is proposed as a new Extended Inter-View Direct mode of macroblock coding and exploits improved inter-view prediction of motion information. For this prediction, the idea is to use depth information in order to predict motion vectors more accurately, as it is possible in the classic Multiview Video Coding. The Extended Inter-View Direct mode is designed especially for the purpose of joint video and depth coding, where depth information for coded view is not available. Extensive experiments show the potential of the new mode to reduce bitstream by 2% to 11% depending on test sequence, compression scenario and variant of the state-of-the-art multiview compression reference technique used.

The new eSBR tool of MPEG-D Universal Speech and Audio Coding offers a great advantage in compression of high frequency content, however it produces audible artifacts for sounds whose pitch frequencies are strongly variable or exceeding the split frequency of eSBR. We propose an extension of the forthcoming standard by adding a high frequency sinusoidal tool. This tool introduces additional parametric information to the data bitstream in order to encode the challenging tonal components which are excluded from eSBR processing. Listening tests demonstrate benefits of the proposed approach for test items of strongly tonal character.

The original idea of the paper is to exploit depth maps in order to increase compression efficiency for multiview video. Depth information is used to establish a 3D mapping between each pixel in an encoded frame and its counterpart in the reference view. With this mapping motion vectors and reference frame indices can be obtained independently for each pixel in a coded picture by a simple derivation of the motion information assigned to the corresponding pixel in the reference view. The goal of the paper is to explore a practical way to use this idea in multiview video coding. This goal has been achieved by experimental testing of various sets of direct and skip modes with and without inter-frame motion prediction. It was shown that either “all modes” have to be used or “all but not classic direct mode” should be used. The respective experimental results have been provided in the paper.

For MPEG-4 AVC/H.264 video encoders, the paper describes a model that defines a relationship between the number of bits for texture information and the quantization step size Q. The model is derived individually for I-, P- and B-frames. The model is defined by formulas given individually for 3 intervals covering whole eligible range of Q values. Extensive experiments with numerous test video sequences have proved good accuracy of the model proposed. In particular the new model exhibits higher accuracy than that from the reference implementation of AVC.

In this paper we propose a new technique for refinement of depth maps. The technique exploits a stereoscopic pair of images and full-pixel precision disparity map in order to produce the output disparity map with sub-pixel precision. The technique employs view synthesis for verification of hypotheses on disparity values, which are formed to iteratively improve the disparity map. Results of experiments, presented for still images and video sequences, show that proposed technique increases quality of depth and has low computational cost. Applications of the proposed technique can be found in future real-time 3D video broadcasting systems which require fast and robust disparity estimation algorithm with high precision.

In this paper we propose a novel approach to error concealment that can be applied to MVC and other 3D video coding technologies. The image content, that is lost due to errors, is recovered with use of multiple error-concealment techniques. In our work we have used three techniques: well-known temporal- and intra-based techniques and a novel inter-view technique. Proposed inter-view recovery employs Depth Image Based Rendering (DIBR), which requires neighboring views and corresponding depth maps. Those depth maps can be delivered in the bit-stream or estimated in the receiver. In order to obtain the final reconstruction, the best technique is selected locally. For that, an original recovery quality measurement method, based on cross-checking, has been proposed. The idea has been implemented and assessed experimentally, with use of 3D video test sequences. The objective and subjective results show that the proposed approach provide good quality of reconstructed video.

Stereo matching techniques usually match segments or blocks of pixels. This paper proposes to match segments defined as fuzzy sets of pixels. The proposed matching method is applicable to various techniques of stereo matching as well as to different measures of differences between pixels. In the paper, embedment of this approach into the state-of-the-art depth estimation software is described. Obtained experimental results show that the proposed way of stereo matching increases reliability of various depth estimation techniques.

Firstly the paper reports experimental results proving that Cascaded Pixel Domain Transcoder (CPDT) is extremely inefficient when used for AVC/H.264 bitstream transcoding aimed at bitrate reduction not exceeding 30% of primary bitrate. In the paper, proposed is a transcoder that exploits structured truncation of bitsream and an algorithm for such truncation is described. The experimental results are described that show that the proposed transcoder provides bitrate reduction with very small or negligible loss of quality for bitrate reductions not exceeding about 30%. The transcoder complexity is much smaller than that of CPDT.

For AVC/H.264 video encoders, the paper describes a simple global model that defines a relationship between the bitrate and the quantization step Qstepfor I-, P- and B-frames. Proposed is a general model for any video shot. As such a model must be dependent on video content, the model parameters have to be estimated from experimental data individually for each video shot, and individually for each frame type. The model is defined by formulas given individually for 3 intervals covering whole eligible range of Qstepvalues. Extensive experiments with numerous test video sequences have proved good accuracy of the model proposed.

Video sequences are compressed by processing individual macroblocks within frames. Traditional macroblock order of processing is linear from the left to the right, and from the top to the bottom. In this paper, an alternative order of macroblock processing is considered. Proposed is the spiral scan beginning at the centre of an image, or more general, in the centre of interest that is often placed in the centre of an image by a person who shots the video. The paper shows that such a change of processing order does not influence compression performance but it is profitable in scalable video coding where it can be used in the enhancement layers by SNR scalability. The paper provides the respective experimental results for low-complexity efficient SNR scalability with fine granularity.

The problem of error accumulation in consecutive compression/decompression cycles of still images is discussed in this article. Considered are predictive and transform techniques together with color transformations. Theoretical analysis is provided for error accumulation in linear transformations followed by quantization. The conditions for lack of error accumulation are discussed. This theoretical analysis is applied to color transformations and compression cycles. Experimental analysis verifies theoretical results and deals with more complicated cases. The experimental analysis includes lossy and near lossless compression as well as color transformations themselves. Moreover, a new technique of near lossless compression with no error accumulation is proposed.

Scalable representation of video mostly consists of some layers that may correspond to various spatial and temporal resolution. In order to match exactly the channel throughput, fine granularity scalability may be used. It is usually achieved by quality (SNR) scalability that scales the bit-stream of an enhancement layer. In this paper, a very simple and straightforward technique of fine granularity scalability is considered. It consists in partitioning of a stream of encoded macroblocks. The new issue is re-ordering of macroblocks according to the spiral scan prior to compression. This re-ordering improves subjective video quality perceived by a viewer that watches scaled video. The advantages of such a technique are: simplicity of the scheme and preservation of the standard bitstream syntax.

Proposed is a coder that produces a layered video representation with layers corresponding to different spatial resolutions. A coder consists of based subcoders with independent motion estimation and compensation. Improved motion-vector encoding is provided for the enhancement layer. Other codec features include adaptive interpolation from the base layer, full AVC-compatibility of the base layer and syntax compatibility of the enhancement-layer bitstream. Interpolated reconstructed base-layer frames are used as additional reference frames in the enhancement layer. The respective prediction modes are embedded into prediction strategy of AVC. Experimental results prove that improved motion-vector encoding results in bitrate reduction of up to 13% of the enhancement-layer bitrate for motion vectors. As measured for all layers together, compression efficiency is significantly higher than for simulcast, but scalable codec complexity is only slightly higher than complexity of the respective simulcast codec.

The paper describes a novel unassisted, fully automatic segmentation technique applicable to colour natural video sequences. The algorithm uses the fast marching method for fast extraction of semantic objects from a frame in a video sequence. These objects are extracted by joint motion and colour analysis. The algorithm handles background in the same way as other objects, thus it does not need global motion compensation. The basic disadvantage of the fast marching method is related to unidirectional motion of an active contour. The new technique overcomes this difficulty by use of an enhancement step consisting in colour processing only. This processing step is performed on a small portion a frame only.

The paper describes a generic multi-loop coder structure suitable for mixed spatial and temporal scalability combined with fine granular SNR scalability. The structure is suitable for various variants of hybrid video coders like MPEG-2, H.263 and H.26L. The idea of mixed spatial and temporal scalability i.e. spatio-temporal scalability is substantial for the proposal. Its application allows improving the scalable coding efficiency i.e. decreasing the scalability overhead. The coder consists of independently motion-compensated sub-coders that produce bitstreams corresponding to individual levels of spatio-temporal resolution. The bitrate can be smoothly matched to the particular channel bandwidth by use of data partitioning, which is related to drift errors in the decoder. Accumulation and propagation of these errors can be bounded by use of proper structures of groups of pictures.

The paper describes a simple empirical model of bitstreams produced by MPEG-2 video coders. The number of bits that represent an I- or P-frame is expressed as a function of the quantization factor Q. This model is global in the sense that it describes the total number of bits in a frame thus ignoring local bitrate variation within a frame as well as changes of Q within a frame. This five-parameter model fits well to experimental data. Four of these parameters may be fixed thus obtaining a less accurate one-parameter model that can be used for coder control. The experimental results are reported for some video test sequences in the BT.601/4CIF format encoded by an MPEG-2 MP@ML coder.

The paper deals with an efficient coder structure being appropriate for scalable coding of video. The coder consists of two motion-compensated hybrid coders with independent motion estimation and compensation. The structure implements spatial scalability or mixed spatial and temporal scalability that can be combined with fine granular SNR scalability. The encoder exhibits extended capabilities of adaptation to network throughput. The H.263 video coding standard is used as a reference but the results are also applicable to the MPEG-2, MPEG-4 and H.26L systems with minor modifications. The coder exhibits a high level of compatibility with standard H.263 and MPEG 2/4 coders.

The paper describes a multi-layer video coder based on spatiotemporal scalability and data partitioning. The coder consists of two parts: a low-resolution coder and full-resolution coder. The first one encodes pictures with reduced spatial and temporal resolution. The data are partitioned into some base layers with fine granularity. The full-resolution encoder exploits interpolated images from the base layers. Here, again the output data can be partitioned into some layers. For the two-layer system, the bitrate overhead measured relative to the single layer MPEG-2 bitstream varies about 10% - 25% for progressive television test sequences. Further layers are related to bitrate overheads of about 3% per layer. The coder structure exhibits high level of compatibility with individual building blocks of MPEG-2 coders.

A comparison of lossless compression results is given for RGB, YC/sub R/C/sub B/ and reversible JPEG 2000 color space. The paper describes the general conditions that rounding errors of a color transformation do not cumulate in the consecutive cycles of forward and inverse transformation. In particular, rounding errors of the RGB/spl rarr/YC/sub R/C/sub B//spl rarr/RGB transformation do not accumulate in the consecutive transformation cycles. Luminance and chrominance representation with two additional bits leads to perfectly reversible color transformation. For such representation, compression is slightly better than in RGB but worse than in JPEG 2000 reversible color space. The paper comprises a study on the accumulation of errors produced by near-lossless JPEG-LS in the consecutive compression-decompression cycles. The paper proves that, alternatively, lossless compression can be performed on luminance and chrominance with reduced representation bit numbers. The advantage is that the errors do not accumulate in the consecutive cycles of compression.

The paper describes video codecs that provide the functionality of spatio-temporal scalability that enables construction of highly scalable video codecs. The coder exploits wavelet decomposition for intraframe coding and combined temporal prediction and spatial interpolation for P- and B-frames. In a two-layer system, the bitrate overhead measured relative to the single layer MPEG-2 bitstream varies about 10% for progressive television test sequences. The base layer bitstream constitutes about 40% of the overall bitstream. In multilayer systems, the bitrate ratio of consecutive layers is about 1:2.5. The base layer encoder is fully compatible with the MPEG-2 video coding standard. The paper comprises experimental results obtained for a two-layer system for progressive video sequence compression.

This paper deals with denoising of color video sequences using median-based filters. It is proved that application of prediction error processing results in improved efficiency of video restoration. The basic idea is to predict a pixel value by using a nonlinear filter, and then to compare this value to that in the input (corrupted) image. Usually these two values are different and a decision has to be made about the pixel value at the filler output. This value can be considered as a sum of the prediction and the prediction error processed in some way. For example, large values of prediction errors are set to zero because they are can be classified as caused by impulsive noise samples. Soft decisions on classification of prediction errors lead to good results for test images.

Recently, there is a great interest in video codecs that implement the functionality of spatial scalability. Unfortunately, those MPEG-2 and MPEG-4 coders which exhibit such a functionality produce much more bits than corresponding single layer coders. This bitrate overhead can be reduced by application of spatio-temporal scalability as proposed by the authors. The base layer bitstream corresponds to pictures with reduced both spatial and temporal resolution while the enhancement layer bitstream is used to transmit the information needed to retrieve images with full spatial and temporal resolution. Full compatibility with the MPEG standards is ensured in the base layer where temporal resolution reduction is obtained by B-frame data partitioning, i.e. by placing each second frame (B-frame) in the enhancement layer only. Improved prediction of B-frames in the enhancement layer is proposed in this paper. The idea is to combine temporal forward and backward prediction with spatial interpolation. Experimental results prove a clear improvement of the MPEG-2-compatible scalable coding efficiency for the scheme proposed.

The goal of the paper is to establish a simple technique for nearly lossless compression of color images which would be resistant to the cumulative effect of distortions introduced in consecutive cycles of encoding and decoding. The paper proves that application of standard color transformation between the RGB and the transmission coordinates leads to maximum error J=1 for the red and green components and d=2 for the blue component. The most important feature is that later cycles of compression and decompression do not introduce any further degradation, so further encoding/decoding cycles remain entirely lossless. Experimental results with histological images have verified the theoretical results.

The existing and standardized solutions for spatial scalability are not satisfactory, therefore new approaches are very actively being explored. The goal of this paper is to improve spatial scalability of MPEG-2 for progressive video. In order to avoid problems with too large bitstreams of the base layer produced by some of the hitherto proposed spatially scalable coders, spatio-temporal scalability is proposed for video compression systems. It is assumed that a coder produces two bitstreams, where the base-layer bitstream corresponds to pictures with reduced both spatial and temporal resolution while the enhancement layer bitstream is used to transmit the information needed to retrieve images with full spatial and temporal resolution. In the base layer, temporal resolution reduction is obtained by B-frame data partitioning, i.e., by placing each second frame (B-frame) in the enhancement layer. Subband (wavelet) analysis is used to provide spatial decomposition of the signal. Full compatibility with the MPEG-2 standard is ensured in the base layer, as compared to single-layer MPEG-2 encoding at bit rates below 6 Mbits/s, the bitrate overhead for scalability is less than 15% in most cases.

The paper deals with design techniques for nonlinear-phase filter banks (FIR and IIR) with applications to image/video subband coding. Nonlinearities of phase characteristics of the analysis filters are compensated in the synthesis filter bank by reversing of data (IIR case) or an impulse response (FIR case). The considerations are restricted to separable multi-level systems, i.e., one-dimensional two-band filter banks are the basic building blocks and therefore carefully considered in the paper. A design technique is described for nonlinear-phase FIR filter banks based on optimization. Moreover the design problem of bireciprocal recursive filters is considered by emphasizing the trade-off between the stopband attenuation and the impulse response length. Both the nonideal separation of subbands and the long impulse responses increase the amount of information to be encoded thus aggravating the overall coding efficiency defined by the relation between the compression ratio and the quality of reconstructed images. The choice of elliptic filters as well as some design techniques improving coding efficiency are also considered.>

Strict definitions of the 2-D filter support, pseudopassivity, and pseudolosslessness are given before new terms of structural pseudopassivity and structural pseudolosslessness are introduced. It is shown that structural pseudopassivity and pseudolosslessness under the finite precision conditions guarantee some useful properties (stability, small sensivity, suppression of parasitic oscillations) of the digital filter but ordinary pseudolosslessness does not. A concise summary of the properties of the structurally pseudolossless 2-D digital filters is given.