In the past decade deep neural networks (DNNs) have shown state-of-the-art performance on a wide range of complex machine learning tasks. Many of these results have been achieved while growing the size of DNNs, creating a demand for efficient compression and transmission of them. In this work we present DeepCABAC, a universal compression algorithm for DNNs that is based on applying Context-based Adaptive Binary Arithmetic Coder (CABAC) to the DNN parameters. CABAC was originally designed for the H.264/AVC video coding standard and became the state-of-the-art for the lossless compression part of video compression. DeepCABAC applies a novel quantization scheme that minimizes a rate-distortion function while simultaneously taking the impact of quantization to the DNN performance into account. Experimental results show that DeepCABAC consistently attains higher compression rates than previously proposed coding techniques for DNN compression. For instance, it is able to compress the VGG16 ImageNet model by x63.6 with no loss of accuracy, thus being able to represent the entire network with merely 9 MB. The source code for encoding and decoding can be found at https://github.com/fraunhoferhhi/DeepCABAC.
9 April 2021 in IEEE Transactions on Circuits and Systems for Video Technology
The paper provides an overview of the quantization and entropy coding methods in the Versatile Video Coding (VVC) standard. Special focus is laid on techniques that improve coding efficiency relative to the methods included in the High Efficiency Video Coding (HEVC) standard: The inclusion of trellis-coded quantization, the advanced context modeling for entropy coding of transform coefficient levels, the arithmetic coding engine with multi-hypothesis probability estimation, and the joint coding of chroma residuals. Beside a description of the design concepts, the paper also discusses motivations and implementation aspects. The effectiveness of the quantization and entropy coding methods specified in VVC is validated by experimental results.
19 January 2011 in Foundations and Trends® in Signal Processing
Digital media technologies have become an integral part of the way we create, communicate, and consume information. At the core of these technologies are source coding methods that are described in this monograph. Based on the fundamentals of information and rate distortion theory, the most relevant techniques used in source coding algorithms are described: entropy coding, quantization as well as predictive and transform coding. The emphasis is put onto algorithms that are also used in video coding, which will be explained in the other part of this two-part monograph.
7 October 2019 in IEEE Transactions on Circuits and Systems for Video Technology
In this paper, we describe a video coding design that enables a higher coding efficiency than the HEVC standard. The proposed video codec follows the design of block-based hybrid video coding, but includes a number of advanced coding tools. A part of the incorporated advanced concepts was developed by the Joint Video Exploration Team, while others are newly proposed. The key aspects of these newly proposed tools are the following. A video frame is subdivided into rectangles of variable size using a binary partitioning with variable split ratios. Three new approaches for generating spatial intra prediction signals are supported: A line-wise application of conventional intra prediction modes, coupled with a mode-dependent processing order, a region-based template matching prediction method and intra prediction modes based on neural networks. For motioncompensated prediction, a multi-hypothesis mode with more than two motion hypotheses can be used. In transform coding, mode dependent combinations of primary and secondary transforms are applied. Moreover, scalar quantization is replaced by trelliscoded quantization and the entropy coding of the quantized transform coefficients is improved. The intra and inter prediction signals can be filtered using an edge-preserving diffusion filter or a non-linear DCT-based thresholding operation. The video codec includes an adaptive in-loop filter for which one of three classifiers can be chosen on a picture basis. We also incorporated an optional encoder control, which adjusts the quantization parameters based on a perceptually motivated distortion measure. In a random access scenario, our proposed video codec achieves luma BD-rate savings between 32.5% for HDR HLG UHD and 39.6% for SDR UHD over the HEVC (HM software) anchor for different categories of test sequences.
1 January 2016 in Foundations and Trends® in Signal Processing
A new method for increasing the coding efficiency of dependent views in multiview video coding by employing already coded motion parameters of other views is presented. The correspondence between a block in a dependent view and a block in a reference view is established by estimating a depth map for the dependent view, either based on already transmitted depth data or by using transmitted disparity vectors and motion parameters. The concept is applicable for multiview coding with and without depth maps. In our experiments with an extension of HEVC, average bit rate savings of about 11% for the dependent views have been measured, which result in approximately 4-5% overall bit rate savings for a 3 view configuration including depth data.