Showing papers on "Lossless compression published in 2020"
TL;DR: Experimental results demonstrate that the coding efficiency of TMC2 is the best on average (especially for lossy geometry and lossy color compression) for dense point clouds while TMC13 achieves the optimal coding performance for sparse and noisy point clouds with lower time complexity.
Abstract: Point cloud based 3D visual representation is becoming popular due to its ability to exhibit the real world in a more comprehensive and immersive way. However, under a limited network bandwidth, it is very challenging to communicate this kind of media due to its huge data volume. Therefore, the MPEG have launched the standardization for point cloud compression (PCC), and proposed three model categories, i.e., TMC1, TMC2, and TMC3. Because the 3D geometry compression methods of TMC1 and TMC3 are similar, TMC1 and TMC3 are further merged into a new platform namely TMC13. In this paper, we first introduce some basic technologies that are usually used in 3D point cloud compression, then review the encoder architectures of these test models in detail, and finally analyze their rate distortion performance as well as complexity quantitatively for different cases (i.e., lossless geometry and lossless color, lossless geometry and lossy color, lossy geometry and lossy color) by using 16 benchmark 3D point clouds that are recommended by MPEG. Experimental results demonstrate that the coding efficiency of TMC2 is the best on average (especially for lossy geometry and lossy color compression) for dense point clouds while TMC13 achieves the optimal coding performance for sparse and noisy point clouds with lower time complexity.
73 citations
TL;DR: This paper proposes to compactly represent and convey the intermediate-layer deep learning features with high generalization capability, to facilitate the collaborating approach between front and cloud ends and makes the standardization of deep feature coding more feasible and promising.
Abstract: The recent advances of hardware technology have made the intelligent analysis equipped at the front-end with deep learning more prevailing and practical. To better enable the intelligent sensing at the front-end, instead of compressing and transmitting visual signals or the ultimately utilized top-layer deep learning features, we propose to compactly represent and convey the intermediate-layer deep learning features with high generalization capability, to facilitate the collaborating approach between front and cloud ends. This strategy enables a good balance among the computational load, transmission load and the generalization ability for cloud servers when deploying the deep neural networks for large scale cloud based visual analysis. Moreover, the presented strategy also makes the standardization of deep feature coding more feasible and promising, as a series of tasks can simultaneously benefit from the transmitted intermediate layer features. We also present the results for evaluations of both lossless and lossy deep feature compression, which provide meaningful investigations and baselines for future research and standardization activities.
71 citations
TL;DR: Experimental results show the superiority of the proposed PW-JND model to the conventional JND models, and a sliding window based search strategy to predict PW- JND based on the prediction results of the perceptually lossy/lossless predictor.
Abstract: Picture Wise Just Noticeable Difference (PW-JND), which accounts for the minimum difference of a picture that human visual system can perceive, can be widely used in perception-oriented image and video processing. However, the conventional Just Noticeable Difference (JND) models calculate the JND threshold for each pixel or sub-band separately, which may not reflect the total masking effect of a picture accurately. In this paper, we propose a deep learning based PW-JND prediction model for image compression. Firstly, we formulate the task of predicting PW-JND as a multi-class classification problem, and propose a framework to transform the multi-class classification problem to a binary classification problem solved by just one binary classifier. Secondly, we construct a deep learning based binary classifier named perceptually lossy/lossless predictor which can predict whether an image is perceptually lossy to another or not. Finally, we propose a sliding window based search strategy to predict PW-JND based on the prediction results of the perceptually lossy/lossless predictor. Experimental results show that the mean accuracy of the perceptually lossy/lossless predictor reaches 92%, and the absolute prediction error of the proposed PW-JND model is 0.79 dB on average, which show the superiority of the proposed PW-JND model to the conventional JND models.
68 citations
TL;DR: In this article, the authors proposed new efficient representations for matrices with low-entropy statistics, which exploit the statistical properties of the data in order to reduce the size and execution complexity.
Abstract: At the core of any inference procedure, deep neural networks are dot product operations, which are the component that requires the highest computational resources. For instance, deep neural networks, such as VGG-16, require up to 15-G operations in order to perform the dot products present in a single forward pass, which results in significant energy consumption and thus limits their use in resource-limited environments, e.g., on embedded devices or smartphones. One common approach to reduce the complexity of the inference is to prune and quantize the weight matrices of the neural network. Usually, this results in matrices whose entropy values are low, as measured relative to the empirical probability mass distribution of its elements. In order to efficiently exploit such matrices, one usually relies on, inter alia, sparse matrix representations. However, most of these common matrix storage formats make strong statistical assumptions about the distribution of the elements; therefore, cannot efficiently represent the entire set of matrices that exhibit low-entropy statistics (thus, the entire set of compressed neural network weight matrices). In this paper, we address this issue and present new efficient representations for matrices with low-entropy statistics. Alike sparse matrix data structures, these formats exploit the statistical properties of the data in order to reduce the size and execution complexity. Moreover, we show that the proposed data structures can not only be regarded as a generalization of sparse formats but are also more energy and time efficient under practically relevant assumptions. Finally, we test the storage requirements and execution performance of the proposed formats on compressed neural networks and compare them to dense and sparse representations. We experimentally show that we are able to attain up to $\times 42$ compression ratios, $\times 5$ speed ups, and $\times 90$ energy savings when we lossless convert the state-of-the-art networks, such as AlexNet, VGG-16, ResNet152, and DenseNet, into the new data structures and benchmark their respective dot product.
58 citations
TL;DR: iWave++ is proposed as a new end-to-end optimized image compression scheme, in which iWave, a trained wavelet-like transform, converts images into coefficients without any information loss, and a single model supports both lossless and lossy compression.
Abstract: Built on deep networks, end-to-end optimized image compression has made impressive progress in the past few years. Previous studies usually adopt a compressive auto-encoder, where the encoder part first converts image into latent features, and then quantizes the features before encoding them into bits. Both the conversion and the quantization incur information loss, resulting in a difficulty to optimally achieve arbitrary compression ratio. We propose iWave++ as a new end-to-end optimized image compression scheme, in which iWave, a trained wavelet-like transform, converts images into coefficients without any information loss. Then the coefficients are optionally quantized and encoded into bits. Different from the previous schemes, iWave++ is versatile: a single model supports both lossless and lossy compression, and also achieves arbitrary compression ratio by simply adjusting the quantization scale. iWave++ also features a carefully designed entropy coding engine to encode the coefficients progressively, and a de-quantization module for lossy compression. Experimental results show that lossy iWave++ achieves state-of-the-art compression efficiency compared with deep network-based methods; on the Kodak dataset, lossy iWave++ leads to 17.34% bits saving over BPG; lossless iWave++ achieves comparable or better performance than FLIF. Our code and models are available at https://github.com/mahaichuan/Versatile-Image-Compression.
57 citations
TL;DR: DeepCABAC as mentioned in this paper applies a novel quantization scheme that minimizes a rate-distortion function while simultaneously taking the impact of quantization to the DNN performance into account, achieving higher compression rates than previously proposed coding techniques for DNN compression.
Abstract: 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 .
54 citations
TL;DR: This article presents a secure, efficient, and complete data collection, and transmission and storage scheme for IoT in smart ocean, able to resist many typical attacks for underwater nodes, such as manipulation attacks, Distributed Denial-of-Service (DDoS) attacks, malicious node injection attacks, and so on.
Abstract: Due to the abundant marine resources, smart ocean has attracted much attention of the government, industry, and academy. The Internet-of-Things (IoT) architectures for smart ocean have been proposed to collect various of data from the ocean, thereby assisting environmental protection, military reconnaissance, and so on. However, few researchers have paid attention to the security and privacy issues of data collection and transmission. In this article, for the unreliable underwater environment, we present a secure, efficient, and complete data collection, and transmission and storage scheme for IoT in smart ocean. Especially, to prolong the lifetime of the underwater node, two novel data compression algorithms [lossy data compression algorithm (LCA) and lossless data compression algorithm (NLCA)] are also proposed. Moreover, due to the vulnerability of underwater nodes, we also propose a corresponding IoT framework and data collection pattern to resist the single point failure attack. Besides, to guarantee the confidentiality, reliability, and integrity of transmitting data, Elliptic Curve-ElGamal (EC-ElGamal) and elliptic curve digital signature algorithm (ECDSA) are employed. The consensus algorithm and blacklisting mechanism are also employed to detect and address failure or malicious nodes. Finally, the security analysis demonstrates that our scheme is able to resist many typical attacks for underwater nodes, such as manipulation attacks, Distributed Denial-of-Service (DDoS) attacks, malicious node injection attacks, and so on. Additionally, relevant experimental results show that the scheme is feasibility and efficiency.
43 citations
TL;DR: The proposed cryptosystem provides the security of the digital images using the cryptographic algorithm in the frequency domain of the image using 2D discrete Haar wavelet transform (DHWT) and 3D logistic chaotic map.
Abstract: This paper proposes a new lossless encryption and decryption method for digital colour-image using 2D discrete Haar wavelet transform (DHWT) and 3D logistic chaotic map. The proposed cryptosystem p...
38 citations
TL;DR: A novel lossless medical image encryption scheme based on game theory with optimized ROI parameters and hidden ROI position achieves optimized and lossless encryption and decryption of images, and can flexibly and reliably protect the medical images of different types and structures against various attacks.
Abstract: Medical images contain a large amount of patients' private information The theft and destruction of medical images will cause irreparable losses to patients and medical institutions In order to detect the region of interest(ROI) accurately, avoid leakage of ROI position information, and realize lossless recovery of transform domain encryption, we propose a novel lossless medical image encryption scheme based on game theory with optimized ROI parameters and hidden ROI position In the encryption process, the ROI is a pixel-level transformed to achieve the lossless decryption of medical images and protect medical image information from loss At the same time, the position information of the ROI is effectively hidden, and leakage of the position information during transmission is avoided In addition, the quantum cell neural network(QCNN) hyperchaotic system generates random sequence to scramble and diffuse the ROI Most important of all, the quantitative analysis method of ROI parameters is given, and the optimal balance between encryption speed and encryption security performance is achieved by using game theory Simulation experiments and numerical analysis verify that the scheme achieves optimized and lossless encryption and decryption of images, and can flexibly and reliably protect the medical images of different types and structures against various attacks
36 citations
TL;DR: It is theoretically proved that a quantum autoencoder can losslessly compress high-dimensional quantum information into a low-dimensional space (also called latent space) if the number of maximum linearly independent vectors from input states is no more than the dimension of the latent space.
Abstract: As a ubiquitous aspect of modern information technology, data compression has a wide range of applications. Therefore, a quantum autoencoder which can compress quantum information into a low-dimensional space is fundamentally important to achieve automatic data compression in the field of quantum information. Such a quantum autoencoder can be implemented through training the parameters of a quantum device using classical optimization algorithms. In this paper, we demonstrate the condition of achieving a perfect quantum autoencoder and theoretically prove that a quantum autoencoder can losslessly compress high-dimensional quantum information into a low-dimensional space (also called latent space) if the number of maximum linearly independent vectors from input states is no more than the dimension of the latent space. Also, we experimentally realize a universal two-qubit unitary gate and design a quantum autoencoder device by applying a machine learning method. Experimental results demonstrate that our quantum autoencoder is able to compress two two-qubit states into two one-qubit states. Besides compressing quantum information, the quantum autoencoder is used to experimentally discriminate two groups of nonorthogonal states.
36 citations
TL;DR: The experimental results show that the proposed lossless coding approach systematically and substantially outperforms the state-of-the-art methods for each type of data.
Abstract: This paper proposes a novel approach for lossless image compression. The proposed coding approach employs a deep-learning-based method to compute the prediction for each pixel, and a context-tree-based bit-plane codec to encode the prediction errors. First, a novel deep learning-based predictor is proposed to estimate the residuals produced by traditional prediction methods. It is shown that the use of a deep-learning paradigm substantially boosts the prediction accuracy compared with the traditional prediction methods. Second, the prediction error is modeled by a context modeling method and encoded using a novel context-tree-based bit-plane codec. Codec profiles performing either one or two coding passes are proposed, trading off complexity for compression performance. The experimental evaluation is carried out on three different types of data: photographic images, lenslet images, and video sequences. The experimental results show that the proposed lossless coding approach systematically and substantially outperforms the state-of-the-art methods for each type of data.
Posted Content•
TL;DR: ResRep as discussed by the authors proposes to re-parameterize a CNN into the remembering parts and forgetting parts, where the former learn to maintain the performance and the latter learn to prune.
Abstract: We propose ResRep, a novel method for lossless channel pruning (a.k.a. filter pruning), which slims down a CNN by reducing the width (number of output channels) of convolutional layers. Inspired by the neurobiology research about the independence of remembering and forgetting, we propose to re-parameterize a CNN into the remembering parts and forgetting parts, where the former learn to maintain the performance and the latter learn to prune. Via training with regular SGD on the former but a novel update rule with penalty gradients on the latter, we realize structured sparsity. Then we equivalently merge the remembering and forgetting parts into the original architecture with narrower layers. In this sense, ResRep can be viewed as a successful application of Structural Re-parameterization. Such a methodology distinguishes ResRep from the traditional learning-based pruning paradigm that applies a penalty on parameters to produce sparsity, which may suppress the parameters essential for the remembering. ResRep slims down a standard ResNet-50 with 76.15% accuracy on ImageNet to a narrower one with only 45% FLOPs and no accuracy drop, which is the first to achieve lossless pruning with such a high compression ratio. The code and models are at this https URL.
TL;DR: The proposed encryption method embeds the encryption into the compression process, in which a small part of the data is encrypted quickly, while maintaining the good coding characteristics of set partitioning in hierarchical trees (SPIHT).
Abstract: In this paper, a novel method for lossless image encryption based on set partitioning in hierarchical trees and cellular automata. The proposed encryption method embeds the encryption into the compression process, in which a small part of the data is encrypted quickly, while maintaining the good coding characteristics of set partitioning in hierarchical trees (SPIHT). The proposed encryption system adopts three stages of scrambling and diffusion. In each stage of encryption, different chaotic systems are used to generate the plaintext-related key stream to maintain high security and to resist some attacks. Moreover, the channel length of the coded-and-compressed color image is more uncertain, resulting into higher difficulty for attackers to decipher the algorithm. The experimental results indicate that the length of bitstream is compressed to 50% of the original image, showing that our proposed algorithm has higher lossless compression ratio compared with the existing algorithms. Meanwhile, the encryption scheme passes the entropy analysis, sensitivity analysis, lossless recovery test, and SP800-22 test.
Posted Content•
TL;DR: A simple and efficient lossless image compression algorithm that predicts the probability of a high-resolution image, conditioned on the low-resolution input, and uses entropy coding to compress this super-resolution operator.
Abstract: We introduce a simple and efficient lossless image compression algorithm. We store a low resolution version of an image as raw pixels, followed by several iterations of lossless super-resolution. For lossless super-resolution, we predict the probability of a high-resolution image, conditioned on the low-resolution input, and use entropy coding to compress this super-resolution operator. Super-Resolution based Compression (SReC) is able to achieve state-of-the-art compression rates with practical runtimes on large datasets. Code is available online at this https URL.
14 Jun 2020
TL;DR: In this article, the authors leverage the powerful lossy image compression algorithm BPG to build a lossless image compression system, where the original image is first decomposed into the lossy reconstruction obtained after compressing it with BPG and the corresponding residual.
Abstract: We leverage the powerful lossy image compression algorithm BPG to build a lossless image compression system. Specifically, the original image is first decomposed into the lossy reconstruction obtained after compressing it with BPG and the corresponding residual. We then model the distribution of the residual with a convolutional neural network-based probabilistic model that is conditioned on the BPG reconstruction, and combine it with entropy coding to losslessly encode the residual. Finally, the image is stored using the concatenation of the bitstreams produced by BPG and the learned residual coder. The resulting compression system achieves state-of-the-art performance in learned lossless full-resolution image compression, outperforming previous learned approaches as well as PNG, WebP, and JPEG2000.
Posted Content•
TL;DR: Modifications to the architecture are introduced to improve the performance of integer discrete flows for lossless compression and discuss the potential difference in flexibility between invertible flows for discrete random variables and flows for continuous random variables.
Abstract: In this paper we analyse and improve integer discrete flows for lossless compression. Integer discrete flows are a recently proposed class of models that learn invertible transformations for integer-valued random variables. Due to its discrete nature, they can be combined in a straightforward manner with entropy coding schemes for lossless compression without the need for bits-back coding. We discuss the potential difference in flexibility between invertible flows for discrete random variables and flows for continuous random variables and show that (integer) discrete flows are more flexible than previously claimed. We furthermore investigate the influence of quantization operators on optimization and gradient bias in integer discrete flows. Finally, we introduce modifications to the architecture to improve the performance of this model class for lossless compression.
TL;DR: Three types of wavelet-based SST (WSST) are obtained by reorganizing all of the existing SSTs covered in this paper by replacing the Haar and Haar-like wavelet transforms with Cohen–Daubechies–Feauveau (CDF) 5/3 and 9/7 wavelets, which are customized on the basis of the original pixel positions in 2D space.
Abstract: Spectral–spatial transforms (SSTs) change a raw camera image captured using a color filter array (CFA-sampled image) from an RGB color space composed of red, green, and blue components into a decorrelated color space, such as YDgCbCr or YDgCoCg color space composed of luma, difference green, and two chroma components. This paper describes three types of wavelet-based SST (WSST) obtained by reorganizing all of the existing SSTs covered in this paper. First, we introduce three types of macropixel SST (MSST) implemented within each $2 \times 2$ macropixel. Next, we focus on two-channel Haar wavelet transforms, which are simple wavelet transforms, and three-channel Haar-like wavelet transforms in each MSST and replace the Haar and Haar-like wavelet transforms with Cohen–Daubechies–Feauveau (CDF) 5/3 and 9/7 wavelet transforms, which are customized on the basis of the original pixel positions in 2D space. Although the test data set is not big, in lossless CFA-sampled image compression based on JPEG 2000, the WSSTs improve the bitrates by about 1.67%–3.17% compared with not using a transform, and the WSSTs that use 5/3 wavelet transforms improve the bitrates by about 0.31%–0.71% compared with the best existing SST. Moreover, in lossy CFA-sampled image compression based on JPEG 2000, the WSSTs show about 2.25–4.40 dB and 26.04%–49.35% in the Bjontegaard metrics (BD-PSNRs and BD-rates) compared with not using a transform, and the WSSTs that use 9/7 wavelet transforms improve the metrics by about 0.13–0.40 dB and 2.27%–4.80% compared with the best existing SST.
TL;DR: The ablation study shows that both DA and the LSA are necessary for high-accuracy face spoofing detection, and the FCN-LSA obtains competitive performance among the state-of-the-art methods.
Abstract: In this paper, a face spoofing detection method called the Fully Convolutional Network with Domain Adaptation and Lossless Size Adaptation (FCN-DA-LSA) is proposed. As its name suggests, the FCN-DA-LSA includes a lossless size adaptation preprocessor followed by an FCN based pixel-level classifier embedded with a domain adaptation layer. The FCN local classifier makes full use of the basic properties of face spoof distortion namely ubiquitous and repetitive. The domain adaptation (DA) layer improves generalization across different domains. The lossless size adaptation (LSA) preserves the high-frequent spoof clues caused by the face recapture process. The ablation study shows that both DA and the LSA are necessary for high-accuracy face spoofing detection. The FCN-LSA obtains competitive performance among the state-of-the-art methods. With the help of small-sample external data in the target domain (2/50, 2/50, and 1/20 subjects for CASIA-FASD, Replay-Attack, and OULU-NPU respectively), the FCN-DA-LSA further improves the performance and outperforms the existing methods.
TL;DR: Numerical simulation results show that the proposed algorithm not only can effectively compression-encryption image, but also have the great security performances, which provides theoretical guide for the application of this algorithm in information safety, and secret communication field.
Abstract: In this paper, a fractional-order memristive band-pass filter (BPF) chaotic circuit is constructed base on BPF chaotic circuit and fractional definition. The attractor and fractal characteristics are analyzed through phase diagrams and time domain response diagrams. In addition, randomness of the chaotic pseudo-random sequences is tested through NIST SP800–22 and correlation of sequence. According to the fractional-order chaotic system and Back-Propagation (BP) neural network, a lossless image compression-encryption algorithm is proposed. In this algorithm, the original image is compressed through BP neural network, and then the compressed image is encrypted by using Zigzag algorithm and xor operation. Numerical simulation results show that the proposed algorithm not only can effectively compression-encryption image, but also have the great security performances, which provides theoretical guide for the application of this algorithm in information safety, and secret communication field.
Posted Content•
TL;DR: The powerful lossy image compression algorithm BPG is leverage to build a lossless image compression system that achieves state-of-the-art performance in learned lossless full-resolution image compression, outperforming previous learned approaches as well as PNG, WebP, and JPEG2000.
Abstract: We leverage the powerful lossy image compression algorithm BPG to build a lossless image compression system. Specifically, the original image is first decomposed into the lossy reconstruction obtained after compressing it with BPG and the corresponding residual. We then model the distribution of the residual with a convolutional neural network-based probabilistic model that is conditioned on the BPG reconstruction, and combine it with entropy coding to losslessly encode the residual. Finally, the image is stored using the concatenation of the bitstreams produced by BPG and the learned residual coder. The resulting compression system achieves state-of-the-art performance in learned lossless full-resolution image compression, outperforming previous learned approaches as well as PNG, WebP, and JPEG2000.
TL;DR: A lossless key-point sequence compression approach for efficient feature coding to eliminate the spatial and temporal redundancies of key points in videos.
Abstract: Feature coding has been recently considered to facilitate intelligent video analysis for urban computing. Instead of raw videos, extracted features in the front-end are encoded and transmitted to the back-end for further processing. In this article, we present a lossless key-point sequence compression approach for efficient feature coding. The essence of this predict-and-encode strategy is to eliminate the spatial and temporal redundancies of key points in videos. Multiple prediction modes with an adaptive mode selection method are proposed to handle key-point sequences with various structures and motion. Experimental results validate the effectiveness of the proposed scheme on four types of widely used key-point sequences in video analysis.
23 Aug 2020
TL;DR: MoSSo as discussed by the authors is the first incremental algorithm for lossless summarization of fully dynamic graphs, which updates the output representation by repeatedly moving nodes among supernodes and edges.
Abstract: Given a fully dynamic graph, represented as a stream of edge insertions and deletions, how can we obtain and incrementally update a lossless summary of its current snapshot? As large-scale graphs are prevalent, concisely representing them is inevitable for efficient storage and analysis. Lossless graph summarization is an effective graph-compression technique with many desirable properties. It aims to compactly represent the input graph as (a) a summary graph consisting of supernodes (i.e., sets of nodes) and superedges (i.e., edges between supernodes), which provide a rough description, and (b) edge corrections which fix errors induced by the rough description. While a number of batch algorithms, suited for static graphs, have been developed for rapid and compact graph summarization, they are highly inefficient in terms of time and space for dynamic graphs, which are common in practice. In this work, we propose MoSSo, the first incremental algorithm for lossless summarization of fully dynamic graphs. In response to each change in the input graph, MoSSo updates the output representation by repeatedly moving nodes among supernodes. MoSSo decides nodes to be moved and their destinations carefully but rapidly based on several novel ideas. Through extensive experiments on 10 real graphs, we show MoSSo is (a) Fast and 'any time': processing each change in near-constant time (less than 0.1 millisecond), up to 7 orders of magnitude faster than running state-of-the-art batch methods, (b) Scalable: summarizing graphs with hundreds of millions of edges, requiring sub-linear memory during the process, and (c) Effective: achieving comparable compression ratios even to state-of-the-art batch methods.
TL;DR: A lower bound on the size of the optimal SPSS is proved and a greedy method called UST is proposed that results in a smaller representation than unitigs and is nearly optimal with respect to the lower bound.
Abstract: Given the popularity and elegance of k-mer based tools, finding a space-efficient way to represent a set of k-mers is important for improving the scalability of bioinformatics analyses. One popular approach is to convert the set of k-mers into the more compact set of unitigs. We generalize this approach and formulate it as the problem of finding a smallest spectrum-preserving string set (SPSS) representation. We show that this problem is equivalent to finding a smallest path cover in a compacted de Bruijn graph. Using this reduction, we prove a lower bound on the size of the optimal SPSS and propose a greedy method called UST that results in a smaller representation than unitigs and is nearly optimal with respect to our lower bound. We demonstrate the usefulness of the SPSS formulation with two applications of UST. The first one is a compression algorithm, UST-Compress, which we show can store a set of k-mers using an order-of-magnitude less disk space than other lossless compression tools. The second one is an exact static k-mer membership index, UST-FM, which we show improves index size by 10-44% compared to other state-of-the-art low memory indices. Our tool is publicly available at: https://github.com/medvedevgroup/UST/.
Posted Content•
TL;DR: This work introduces an algorithm that removes units and layers of a neural network while not changing the output that is produced, which thus implies a lossless compression.
Abstract: Deep neural networks have been successful in many predictive modeling tasks, such as image and language recognition, where large neural networks are often used to obtain good accuracy. Consequently, it is challenging to deploy these networks under limited computational resources, such as in mobile devices. In this work, we introduce an algorithm that removes units and layers of a neural network while not changing the output that is produced, which thus implies a lossless compression. This algorithm, which we denote as LEO (Lossless Expressiveness Optimization), relies on Mixed-Integer Linear Programming (MILP) to identify Rectified Linear Units (ReLUs) with linear behavior over the input domain. By using L1 regularization to induce such behavior, we can benefit from training over a larger architecture than we would later use in the environment where the trained neural network is deployed.
TL;DR: To the authors' knowledge, the paper is the first to replace all the traditional HEVC-based angular intra-prediction modes with an intra-Prediction method based on modern Machine Learning techniques for lossless video coding applications.
Abstract: The paper proposes a novel block-wise prediction paradigm based on Convolutional Neural Networks (CNNs) for lossless video coding. A deep neural network model which follows a multi-resolution design is employed for block-wise prediction. Several contributions are proposed to improve neural network training. A first contribution proposes a novel loss function formulation for an efficient network training based on a new approach for patch selection. Another contribution consists in replacing all HEVC-based angular intra-prediction modes with a CNN-based intra-prediction method, where each angular prediction mode is complemented by a CNN-based prediction mode using a specifically trained model. Another contribution consists in an efficient adaptation of the CNN-based intra-prediction residual for lossless video coding. Experimental results on standard test sequences show that the proposed coding system outperforms the HEVC standard with an average bitrate improvement of around 5%. To our knowledge, the paper is the first to replace all the traditional HEVC-based angular intra-prediction modes with an intra-prediction method based on modern Machine Learning techniques for lossless video coding applications.
TL;DR: This letter presents a lossless intra coder of the geometry information of voxelized point clouds that outperforms all state-of-the-art intra coders on the public available point cloud datasets tested.
Abstract: This letter presents a lossless intra coder of the geometry information of voxelized point clouds. Instead of using the popular octree decomposition, the proposed method views the point cloud geometry as an array of bi-level images, and it is inspired by well-known techniques for coding this type of images. This array is encoded using a dyadic decomposition that recursively splits the array into two arrays of half its size, transmitting the occupancy information of each smaller array. Context adaptive arithmetic coding, using both 2D and 3D contexts, is used to achieve efficient compression. Results show that the proposed method outperforms all state-of-the-art intra coders on the public available point cloud datasets tested.
Proceedings Article•
30 Apr 2020TL;DR: In this article, the authors make the following striking observation: fully convolutional VAE models trained on 32x32 ImageNet can generalize well to 64x64 and far larger photographs, with no changes to the model.
Abstract: We make the following striking observation: fully convolutional VAE models trained on 32x32 ImageNet can generalize well, not just to 64x64 but also to far larger photographs, with no changes to the model. We use this property, applying fully convolutional models to lossless compression, demonstrating a method to scale the VAE-based 'Bits-Back with ANS' algorithm for lossless compression to large color photographs, and achieving state of the art for compression of full size ImageNet images. We release Craystack, an open source library for convenient prototyping of lossless compression using probabilistic models, along with full implementations of all of our compression results.
TL;DR: The design and implementation of a lossless/near-lossless compression system for high-frame-rate gaze camera image data, which is faced with the technical problems of high fidelity and strong real-time and reliable compression.
Abstract: This paper explores dynamic visual communication image framing for graphic design based on virtual reality algorithms; it defines corresponding feature representations by delineating layers of pixels, elements, relationships, planes, and applications; and it investigates methods for quantifying geometric features, perceptual features, and style features. The contents include extraction methods for element colors, calculation methods for layout perceptual features and color-matching perceptual features, and pairwise comparison methods for style features. By overfitting the distribution of geometric features in the data, the model can predict the probability density distribution of features such as element position and color under specific conditions to support the generation of flat images. To construct a prediction model, the sampling method of features, the model optimization method, and the data learning strategy are investigated. This thesis involves the design and implementation of a lossless/near-lossless compression system for high-frame-rate gaze camera image data, which is faced with the technical problems of high fidelity and strong real-time and reliable compression. The image single-frame lossless/near-loss-free compression ratio is generally low, and the compression ratio can be improved by using the correlation between image frames. In this paper, we study the application of lossless compression between image frames, the efficient computing structure of FPGA, and an onboard compression system.
TL;DR: The design, implementation and results of a set of IP cores that perform on- board hyperspectral image compression according to the CCSDS 123.0-B-1 lossless standard are presented, specifically designed to be suited for on-board systems and for any kind of hyperspectrals sensor.
Abstract: In this paper, we present the design, implementation and results of a set of IP cores that perform on-board hyperspectral image compression according to the CCSDS 123.0-B-1 lossless standard, specifically designed to be suited for on-board systems and for any kind of hyperspectral sensor. As entropy coder, the sample-adaptive entropy coder defined in the 123.0-B-1 standard or the low-complexity block-adaptive encoder defined by the CCSDS 121.0-B-2 lossless standard could be used. Both IPs, 123.0-B-1 and 121.0-B-2, are part of SHyLoC 2.0, and can be used together for compression of hyperspectral images, being also possible the compression of any kind of data using only the 121-IP. SHyLoC 2.0 improves and extends the capabilities of SHyLoC 1.0, currently available at the ESA IP Cores library, increasing its compression efficiency and throughput, without compromising the resources footprint. Moreover, it incorporates new features, such as the unit-delay predictor option defined by the CCSDS 121.0-B-2 standard, and burst capabilities in the external memory interface of the CCSDS 123-IP, among others. Dedicated architectures have been designed for all the possible input image sample arrangements, in order to maximise throughput and reduce the hardware resources utilization. The design is technology-agnostic, enabling the mapping of the VHDL code in different FPGAs or ASICs. Results are presented for a representative group of well-known space-qualified FPGAs, including the new NanoXplore BRAVE family. A maximum throughput of 150 MSamples/s is obtained for Xilinx Virtex XQR5VFX130 when the SHyLoC 2.0 CCSDS-123 IP is configured in Band-Interleaved by Pixel (BIP) order, using only the 4% of LUTs and less than the 1% of internal memory.
TL;DR: A lossless lightweight design strategy is proposed for CNN to efficiently achieve the SAR target recognition, which subtly utilizes pruning and knowledge distillation.
Abstract: Due to high computational cost and large memory overhead, it is difficult to deploy original deep convolutional neural network (CNN) on real-time embedded devices of synthetic aperture rada...