Maciej Bartkowiak

Enhanced coding of high-frequency tonal components in MPEG-D USAC through joint application of ESBR and sinusoidal modeling

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.

New results in free-viewpoint television systems for horizontal virtual navigation

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.

Switched-current filter design for image processing systems

The paper presents a design method of two-dimensional. (2-D) switched-current (SI) filters. Because of SI technique compatibility with the standard digital CMOS technology, it is very convenient to use such filters in image processing. Operation of a system containing 2-D filters in a pre-processing stage is described. Such a system can be implemented in portable one-chip devices like digital video cameras.

Interpolation of nonuniformly sampled chrominance using luminance information

The paper deals with a new interpolation technique for reconstruction of nonuniformly and irregularly sampled color images. The assumption is that there is at least one densely sampled component known (the luminance component). The presented technique is a simple noniterative adaptive approach that exploits intercomponent correlation of natural scene images. The approach is based on adaptive adjustment of triangle interpolation coefficients. A simple adaptation rule is derived from analysis of gradient magnitudes that exhibit high correlation between color components in natural scene images. This scenario is employed in very low bit rate coding of color images with nonuniform sampling applied to chrominance components.

Nonuniform sampling of chrominance and its application to intra-frame coding

The paper deals with signal-dependent nonuniform sampling of chrominance. Correlation between the luminance and both chrominance components is exploited to estimate the sample position. In this way, the need for side information is avoided. These results are used in a new technique of intraframe coding of chrominance. Usually, in the very low bitrate interframe coding of video, very few bits are allocated to chrominance. Therefore the quality of chrominance in the decoded I-frames has a great influence on the overall chrominance quality in a senes of consecutive P-frames. The paper describes two variants of a new technique for compression of chrominance data in color images and video sequences coded at low bitrates.

Efficient Representation of Chrominance for Very Low Bitrate Coding

The paper describes an original method to represent chrominance in color images and video. This method can be combined with an arbitrary technique of luminance representation or compression. The two chrominance components are represented by one scalar signal obtained using vector quantization. The luminance is encoded entirely independently from chrominance. The scalar representation of chrominance exhibits high redundancy and high correlation with luminance. They both are strongly reduced using differential coding with adaptive prediction that exploits the information about edges extracted from decoded luminance component.

A Simple Algorithm for Ordering and Compression of Vector Codebooks

The problem of storage or transmission of codevectors is an essential issue in vector quantization with custom codebook. The proposed technique for compression of codebooks relies on structuring and ordering properties of a binary split algorithm used for codebook design. A simple algorithm is presented for automatic ordering of the codebook entries in order to group similar codevectors. This similarity is exploited in efficient compression of the codebook content by the means of lossless differential coding and lossy DCT-based coding. Experimental results of two compression experiments are reported and show that a small compression gain can be achieved in this way.

Color representation using scalar chrominance

The paper deals with color image and video representation that consists of two components instead of three. The first component is luminance defined in a usual way. The second component is scalar chrominance obtained from two chrominance components to one scalar chrominance is a vector quantization task which can be performed quite efficiently using a binary split algorithm. Experimental results for color images in CIF and QCIF resolution prove that application of a codebook with 20-60 different scalar chrominances leads to a representation that is mostly indistinguishable from the original images. Since the sets of scalar chrominance values are small, scalar chrominance samples need only short binary representations. Moreover scalar chrominance can be subsampled as chrominance usually is. The mapping of the set of chrominance value pairs on a subset of integer numbers defines an order in the codebook. This order is also an important issue as it influences spectrum of the picture of scalar chrominance and exhibits substantial impact on further compression and processing of scalar chrominance.

Chrominance Vector Quantization for Coding of Images and Video at Very Low Bitrates

Publisher Summary Vector quantization (VQ) is well-known as a powerful technique for image and video data compression. There are many practical possibilities to create vectors in an image. VQ, where vector components are the luminances of some neighboring pixels, is often used. Another approach is to use the color coordinates of a pixel as the vector value which is then quantized. The chapter describes VQ in the two-dimensional space of the chrominance coordinates U and V. This VQ results in one scalar signal which is then processed independent of the luminance in any coder. The coder processes only one chrominance component instead of two. Proper ordering of the codebook decreases the frequency band of this scalar chrominance signal. The chapter also describes an efficient and non-iterative technique to generate the code book which is used to encode the chrominance pairs.