Recent approaches to compression of deep neural networks, like the emerging standard on compression of neural networks for multimedia content description and analysis (MPEG-7 part 17), apply scalar quantization and en...
详细信息
ISBN:
(纸本)9781728163956
Recent approaches to compression of deep neural networks, like the emerging standard on compression of neural networks for multimedia content description and analysis (MPEG-7 part 17), apply scalar quantization and entropy coding of the quantization indexes. In this paper we present an advanced method for quantization of neural network parameters, which applies dependent scalar quantization (DQ) or trellis-coded quantization (TCQ), and an improved context modeling for the entropy coding of the quantization indexes. We show that the proposed method achieves 5.778% bitrate reduction and virtually no loss (0.37%) of network performance in average, compared to the baseline methods of the second test model (NCTM) of MPEG-7 part 17 for relevant working points.
The purpose of the research work presented here is to reduce the execution time of arithmetic coding. arithmetic coding is a very efficient and most popular entropy coding technique used with most of the data compress...
详细信息
ISBN:
(纸本)9781479901920
The purpose of the research work presented here is to reduce the execution time of arithmetic coding. arithmetic coding is a very efficient and most popular entropy coding technique used with most of the data compression methods. arithmetic coding is based on cumulative probability of symbols. In implementations using integer arithmetic, it uses cumulative frequency/total frequency in place of cumulative probability while computing new subinterval [low, high). In renormalization loop, it outputs single matching most significant bit of low and high bound of an interval and renormalizes an interval till an interval becomes 2(b-2) wide, where b is number of bits used to store the range value. In this paper, multibit-power2 approach is proposed that combines two techniques together to increase the speed of execution. One technique is to scale all frequencies such that total frequency results in power of two. This enables division or multiplication by total frequency to be performed using shift operations. Another technique is to process more than one bit at a time to reduce the number of iterations of renormalization loop. As compared to conventional implementations, proposed multibit-power2 approach has resulted in 82% overall saving in the execution speed while compressing with no compromise on compression rate and 38% overall gain in execution time while decompressing.
We continue developing the information theory of structured data. In this article, we study models generating d-ary trees (d >= 2) and trees with unrestricted degree. We first compute the entropy which gives us the...
详细信息
We continue developing the information theory of structured data. In this article, we study models generating d-ary trees (d >= 2) and trees with unrestricted degree. We first compute the entropy which gives us the fundamental lower bound on compression of such trees. Then we present efficient compression algorithms based on arithmetic encoding that achieve the entropy within a constant number of bits. A naive implementation of these algorithms has a prohibitive time complexity of O(n(d)) elementary arithmetic operations (each corresponding to a number f (n(d)) of bit operations), but our efficient algorithms run in O(n(2)) of these operations, where n is the number of nodes. It turns out that extending source coding (i.e., compression) from sequences to advanced data structures such as degree-unconstrained trees is mathematically quite challenging and leads to recurrences that find ample applications in the information theory of general structures (e.g., to analyze the information content of degree-unconstrained non-plane trees).
A new Modified Discrete Wavelets Packets Transform (MDWPT) based method for the compression of Surface EMG signal (s-EMG) data is presented. A Modified Discrete Wavelets Packets Transform (MDWPT) is applied to the dig...
详细信息
A new Modified Discrete Wavelets Packets Transform (MDWPT) based method for the compression of Surface EMG signal (s-EMG) data is presented. A Modified Discrete Wavelets Packets Transform (MDWPT) is applied to the digitized s-EMG signal. A Discrete Cosine Transforms (DCT) is applied to the MDWPT coefficients (only on detail coefficients). The MDWPT+ DCT coefficients are quantized with a Uniform Scalar Dead-Zone Quantizer (USDZQ). An arithmetic coder is employed for the entropy coding of symbol streams. The proposed approach was tested on more than 35 actuals S-EMG signals divided into three categories. The proposed approach was evaluated by the following parameters: Compression Factor (CF), Signal to Noise Ratio (SNR), Percent Root mean square Difference (PRD), Mean Frequency Distortion (MFD) and the Mean Square Error (MSE). Simulation results show that the proposed coding algorithm outperforms some recently developed s-EMG compression algorithms.
In the present world data compression is used in every field. Through data compression the bits required to represent a message will be reduced. By compressing the given data, we can save the storage capacity, files a...
详细信息
In the present world data compression is used in every field. Through data compression the bits required to represent a message will be reduced. By compressing the given data, we can save the storage capacity, files are transferred at high speed, storage hardware is decreased so that its cost is also decreased, and storage bandwidth is decreased. There are many methods to compress the data. But in this paper, we are discussing about Huffman coding and arithmetic coding. For various input streams we are comparing adaptive Huffman coding and arithmetic coding and we will observe which technique will be more efficient to compress the data.
Introduction of processor based instruments in power systems is resulting in the rapid growth of the measured data volume. The present practice in most of the utilities is to store only some of the important data in a...
详细信息
Introduction of processor based instruments in power systems is resulting in the rapid growth of the measured data volume. The present practice in most of the utilities is to store only some of the important data in a retrievable fashion for a limited period. Subsequently even this data is either deleted or stored in some back up devices. The investigations presented here explore the application of lossless data compression techniques for the purpose of archiving all the operational data - so that they can be put to more effective use. Four arithmetic coding methods suitably modified for handling power system steady state operational data are proposed here. The performance of the proposed methods are evaluated using actual data pertaining to the Southern Regional Grid of India. (C) 2012 Elsevier Ltd. All rights reserved.
Medical imaging in hospitals requires fast and efficient image compression to support the clinical work flow and to save costs. Least-squares autoregressive pixel prediction methods combined with arithmetic coding con...
详细信息
ISBN:
(纸本)9781479923427
Medical imaging in hospitals requires fast and efficient image compression to support the clinical work flow and to save costs. Least-squares autoregressive pixel prediction methods combined with arithmetic coding constitutes the state of the art in lossless image compression. However, a high computational complexity of both prevents the application of respective CPU implementations in practice. We present a massively parallel compression system for medical volume images which runs on graphics cards. Image blocks are processed independently by separate processing threads. After pixel prediction with specialized border treatment, prediction errors are entropy coded with an adaptive binary arithmetic coder. Both steps are designed to match particular demands of the parallel hardware architecture. Comparisons with current image and video coders show efficiency gains of 3.3-13.6% while compression times can be reduced to a few seconds.
Aiming to improve the efficiency of image compression on the premise of controlling the compression time, an image compression oriented algorithm is proposed based on Dual Tree-Complex Wavelet and arithmetic coding al...
详细信息
ISBN:
(纸本)9781728150451
Aiming to improve the efficiency of image compression on the premise of controlling the compression time, an image compression oriented algorithm is proposed based on Dual Tree-Complex Wavelet and arithmetic coding algorithm. Methods : decompose the read-in image by Dt-cwt, then using Wellner (an adaptive binarization method) to calculate the image's threshold. Scan the image by EZW algorithm,and compared the threshold with wavelet coefficients,then output the important coefficients and compress these data by arithmetic coding. Finally, the compressed image was output by encoding inverse transform. Results :Four-level decomposition of Dual Tree-Complex Wavelet and seven-times scanning of the image have better compression result and shorter running time. Conclusion :The hardware configuration of this method is simple, and this algorithm can compress the image with high quality in a short time.
It seems reasonable to expect from a good compression method that its output should not be further compressible, because it should behave essentially like random data. We investigate this premise for a variety of know...
详细信息
It seems reasonable to expect from a good compression method that its output should not be further compressible, because it should behave essentially like random data. We investigate this premise for a variety of known lossless compression techniques, and find that, surprisingly, there is much variability in the randomness, depending on the chosen method. arithmetic coding seems to produce perfectly random output, whereas that of Huffman or Ziv-Lempel coding still contains many dependencies. In particular, the output of Huffman coding has already been proven to be random under certain conditions, and we present evidence here that arithmetic coding may produce an output that is identical to that of Huffman.
Linguistic steganography aims to hide secret messages within text carriers. In this paper, we propose a linguistic steganography method by means of sampling-based language generation. Comparing with deterministic text...
详细信息
ISBN:
(纸本)9781728132488
Linguistic steganography aims to hide secret messages within text carriers. In this paper, we propose a linguistic steganography method by means of sampling-based language generation. Comparing with deterministic text generation using beam-search, the sampling-based approach increases the redundancy of generated texts and benefits the hiding of information. The arithmetic coding (AC) algorithm is adopted to embed messages in our proposed method. Its performance is compared with fixed-length coding (FLC) and variable-length coding (VLC) which were designed for embedding messages during deterministic text generation. Besides, the KL divergence and temperature based strategies are designed to control the embedding rates of FLC, VLC and AC respectively. Experiments using a story generation model show that AC performed better than FLC and VLC when embedding messages during sampling-based text generation. With an embedding rate of 1.45 bits/word, our AC-based steganography method achieved ideal imperceptibility, and the subjective quality of its generated text is as good as the non-steganography one.
暂无评论