Scalable Image Compression Research Articles

Exploring Resolution Fields for Scalable Image Compression With Uncertainty Guidance

Exploring Resolution Fields for Scalable Image Compression With Uncertainty Guidance

Scalable image compression mechanism for surveillance video summary

Scalable image compression mechanism for surveillance video summary

Adaptive Channel Scheduling for Acceleration and Fine Control of RNN-Based Image Compression

The existing target-dependent scalable image compression network can control the target of the compressed images between the human visual system and the deep learning based classification task. However, in its RNN based structure controls the bit-rate through the number of iterations, where each iteration generates a fixed size of the bit stream. Therefore, a large number of iterations are required at the high BPP, and fine-grained image quality control is not supported at the low BPP. In this paper, we propose a novel RNN-based image compression model that can schedule the channel size per iteration, to reduce the number of iterations at the high BPP and fine-grained bit-rate control at the low BPP. To further enhance the efficiency, multiple network models for various channel sizes are combined into a single model using the slimmable network architecture. The experimental results show that the proposed method achieves comparable performance to the existing method with finer BPP adjustment, increases parameters by only 0.15% and reduces the average amount of computation by 40.4%.

BlinQS: Blind quality scalable image compression algorithm without using PCRD optimization

BlinQS: Blind quality scalable image compression algorithm without using PCRD optimization

Scalable image compression algorithms with small and fixed-size memory

Scalable image compression algorithms with small and fixed-size memory

Scalable compression for machine and human vision tasks via multi-branch shared module

The existing image compression methods are mainly aimed at human perception tasks instead of machine vision tasks. Rich features learned by the shallow and deep layers of pre-trained visual geometry group (VGG)-net can serve human perception and machine vision tasks, respectively. To improve the machine analysis capabilities of the human-targeted compression method, we propose a scalable image compression framework at low bit-rates. Specifically, the scalable compression framework is composed of a base layer (BL) and an enhancement layer (EL), which utilize the correlation between the above features to perform machine analysis and human perception tasks, respectively. For effectively utilizing the above two types of features, a multi-branch shared module that utilizes the complementarity of multi-branch convolution kernels to retain compact information for the above features to support BL and EL tasks is proposed. In addition, for further improving the accuracy of machine analysis, the machine-vision importance map is introduced in BL; it adaptively utilizes the spatial and channel information from the deepest layer of VGG-net to guide local bit allocation. When the bit-rate is limited to 0.2 bpp, the average recognition accuracy (Top-1) of the BL of proposed method is 5.2%, 13.4%, 6.0%, and 7.4% higher than that of BPG, Webp, Mentzer, and NIC, respectively, on the ILSVRC2012 verification dataset. Meanwhile, the EL provides good visual experience.

Learning-Based Scalable Image Compression With Latent-Feature Reuse and Prediction

Recently, learning-based image compression model has attracted much attention due to its impressive performance and ease of optimization, compared with traditional DCT and wavelet-based image compression standards. Most learning-based image compression models are trained to minimize joint rate-distortion (RD) loss on one single RD trade-off point. However, in many multimedia applications, due to communication constraints, or display adaptation needs for different spatial formats, bit rates or power, it is necessary to provide a variety of image versions for different client devices. To fulfill this requirement, typical end-to-end image compression methods have to compress an image into several bit streams independently by a number of pre-trained networks, which are resource-consuming because of redundancy among these streams. To address this problem, inspired by traditional scalable video coding framework, we propose a learning-based end-to-end quality and spatial scalable image compression (QSSIC) model in multi-layer structure, in which each layer could generate one bitstream corresponding to a specified resolution and image fidelity. This scalability is achieved by exploring the potential of feature-domain representation prediction and reuse. To be specific, firstly, bitstreams of previous layers are used to predict the current layer representations which contains the enhancement information, and then only prediction residuals need to be coded in enhancement layers. Secondly, previous bitstreams are reused in image reconstruction in higher layers to provide basic information. The proposed model could be optimized in an end-to-end manner. Extensive experiments show that our method outperforms state-of-art deep neural networks (DNN)-based auto-encoders in simulcast scenarios. In addition, our method has a better performance than the traditional scalable image compression method scalable extension of H.264/AVC (SVC) and is comparable to scalable extension of H.265/HEVC (SHVC).

Scalable image compression mechanism for surveillance video summary

Scalable image compression mechanism for surveillance video summary

Semantics-to-Signal Scalable Image Compression with Learned Revertible Representations

Semantics-to-Signal Scalable Image Compression with Learned Revertible Representations

Target-Dependent Scalable Image Compression Using a Reconfigurable Recurrent Neural Network

Conventional human-centric image compression techniques are optimized for human visual perception, and are generally evaluated by metrics such as MSSSIM and PSNR. On the other hand, task-centric image compression techniques that target deep neural networks (DNN) inference focus on understanding images, and are measured by inference accuracy. As images should be encoded and decoded differently depending on the target metric, existing deep learning based compression techniques have focused on designing an independent neural network that is optimized for either one of the two targets. However, as these techniques require the target metric to be determined during a training phase, separate DNN models should be trained for different target applications, allowing no scalability between human- and task-centric compression. In this paper, we propose a target-dependent scalable image compression pipeline, where compressed images can be differently decoded in real time depending on the interpolated target between perceptual quality and DNN inference accuracy. This dynamic scalability is supported by a reconfigurable recurrent neural network (RNN) that can dynamically change its flow according to the interpolation parameter. The experimental results show that the proposed method achieves comparable MSSSIM and classification accuracy to the compression networks individually trained for human-centric and task-centric targets, respectively, while providing scalability between the two targets depending on the interpolation parameter.

Efficient and simple scalable image compression algorithms

Abstract Most wavelet-based scalable image compression algorithms are optimized for performance only, i.e., they try to produce a compressed bit-stream that gives the best image quality for the selected bit-rate. Unfortunately, the optimization of the coding method may result in a bit-stream that cannot be rearranged easily to support resolution scalability. This paper proposes two algorithms. The first algorithm produces a rate scalable bit-stream that is optimized for performance and scalability. That is, it has good performance and at the same time it produces bit-stream that can be easily rearranged to become highly (rate and resolution) scalable. To this end, the algorithm combines the low complexity and rate scalability features of the set partitioning coding approach, and it exploits the resolution scalability characteristics of the block-based systems. The second algorithm upgrades the first algorithm to produce a highly scalable bit-stream by only arranging and identifying the different resolutions within the bit-stream.

Efficient Scalable Image Compression Algorithms with Low Memory and Complexity

The set partitioning embedded block (SPECK) image compression algorithm has excellent performance, low computational complexity, and produces a rate scalable compressed bitstream that can be decoded efficiently at multiple bit-rates. Unfortunately, it consumes a huge amount of computer memory due to employing lists that store the coordinates of the image pixels and the coordinates of the sets that are generated during the coding process. In addition, it has complex memory management due to using an array of random access linked lists to store these sets according to their sizes. In this paper, we propose two algorithms that are based on SPECK. The main contribution of the first algorithm is that, as compared to SPECK, the amount of the algorithm’s usable memory is reduced to about 75% and at the same time its processing speed is increased and its rate distortion efficiency is preserved as will be demonstrated. The second algorithm has higher processing speed but has slightly lower rate distortion performance than the first algorithm.

Highly Scalable Single List Set Partitioning in Hierarchical Trees Image Compression

A highly scalable image compression system produces a compressed image that can be easily decoded at several bit-rates (qualities) and resolutions (sizes). This is a very interesting feature in the heterogeneous environments of the Internet to satisfy the different user's requests. This paper proposes a new algorithm termed Highly Scalable-Single List-SPIHT (HS-SLS). The proposed HS-SLS algorithm produces a highly scalable bit-stream that is both rate and resolution scalable. The flexible bit-stream of the HS-SLS encoder can easily be adapted to various resolution requirements at any bit-rate using a very simple scaling process that is performed on-the-fly and without the need to decode the bit-stream. A unique feature of the new algorithm is that it has low memory requirement and low complexity due to its simple memory management. In addition, the size of the usable memory is fixed and thus can be predefined which avoids the dynamic memory allocation problem. These features make the HS-SLS algorithm very suitable for hardware implementation. The price paid for these valuable features is very slight decrement in the algorithm's performance as compared to the original SPIHT algorithm. Index Terms: Highly Scalable Image Compression, Low Memory Image Compression, Rate scalability, Resolution scalability, Set Partitioning image Compression, SPIHT, Wavelet image compression, Zero-tree coding.

ANALYSIS OF IMAGE COMPRESSION ALGORITHMS USING WAVLET TRANSFORM WITH GUI IN MATLAB

Image compression is nothing but reducing the amount of data required to represent an image. To compress an image efficiently we use various techniques to decrease the space and to increase the efficiency of transfer of the images over network for better access. This paper explains about compression methods such as JPEG 2000, EZW, SPIHT (Set Partition in Hierarchical Trees) and HSSPIHT on the basis of processing time, error comparison, mean square error, peak signal to noise ratio and compression ratio. Due to the large requirement for memory and the high complexity of computation, JPEG2000 cannot be used in many conditions especially in the memory constraint case. SPIHT gives better simplicity and better compression compared to the other techniques. But to scale the image more so as to get better compression we are using the line-based Wavelet transform because it requires lower memory without affecting the result of Wavelet transform. We proposed a highly scalable image compression scheme based on the Set Partitioning in Hierarchical Trees (SPIHT) algorithm. This algorithm is called Highly Scalable SPIHT (HS_SPIHT) it gives good scalability and provides 1 bit stream that can be easily adapted to give bandwidth and resolution requirements.

A New Method of Scalable Image Compression Coding

This paper presents a new method of scalable image compression coding based on the wavelet transform. This method delimits the region of interest of the original image, and give a high-quality encoding to this region and a rough encoding to the rest. The result shows that in the limited memory space, this algorithm provides the coarser reconstruct image to satisfy the basal subject quality. Using this method, we can give a high quality encoding to the region for interest.

A Novel VLSI Architecture for SPHIT Encoder

In this Paper we propose a highly scalable image compression scheme based on the set partitioning in hierarchical trees (SPIHT) algorithm. Our algorithm called highly scalable SPIHT (HS-SPIHT), supports spatial and SNR scalability and provides a bit stream that can be easily adapted (reordered) to given bandwidth and resolution requirements by a simple transcoder (parser). The HS-SPIHT algorithm adds the spatial scalability feature without sacrificing the SNR embeddedness property as found in the original SPIHT bit stream. HS-SPIHT finds applications in progressive Web browsing, flexible image storage and retrieval, and image transmission over heterogeneous networks. Here we have written the core processor Microblaze is designed in VHDL (VHSIC hardware description language), implemented using XILINX ISE 8.1 Design suite the algorithm is written in system C Language and tested in SPARTAN-3 FPGA kit by interfacing a test circuit with the PC using the RS232 cable. The test results are seen to be satisfactory. The area taken and the speed of the algorithm are also evaluated.

Adaptive Transmission of Medical Image and Video Using Scalable Coding and Context-Aware Wireless Medical Networks

The aim of this paper is to present a novel platform for advanced transmission of medical image and video, introducing context awareness in telemedicine systems. Proper scalable image and video compression schemes are applied to the content according to environmental properties (i.e., the underlying network status, content type, and the patient status). The transmission of medical images and video for telemedicine purposes is optimized since better content delivery is achieved even in the case of low-bandwidth networks. An evaluation platform has been developed based on scalable wavelet compression with region-of-interest support for images and adaptive H.264 coding for video. Corresponding results of content transmission over wireless networks (i.e., IEEE 802.11e, WiMAX, and UMTS) have proved the effectiveness and efficiency of the platform.

Flexible, highly scalable, object-based wavelet image compression algorithm for network applications

The authors propose a highly scalable image compression scheme based on the set partitioning in hierarchical trees (SPIHT) algorithm. The proposed algorithm, called highly scalable SPIHT (HS-SPIHT), adds the spatial scalability feature to the SPIHT algorithm through the introduction of multiple resolution-dependent lists and a resolution-dependent sorting pass. It keeps the important features of the original SPIHT algorithm such as compression efficiency, full SNR scalability and low complexity. The flexible bitstream of the HS-SPIHT encoder can easily be adapted to various resolution requirements at any bit rate. The parsing process can be carried out on-the-fly without decoding the bitstream by a simple parser (transcoder) that forms a part of a smart network. The HS-SPIHT algorithm is further developed for fully scalable coding of arbitrarily shaped visual objects. The proposed highly scalable algorithm finds applications in progressive web browsing, visual databases and especially in image transmission over heterogeneous networks.

Image Compression And Watermarking By Wavelet Localization

This paper has presented two image processing applications, compression and watermarking, by exploiting the localization property of wavelet transform. The first application proposes a simple region-based and scalable image compression algorithm. We locate the wavelet coefficients in the region of interest in each subband, and these groups of wavelet coefficients are used to adjust the resolution of the interested region. A watermarking method is described in the second application of this paper. The scheme examines the variations of the local Hölder regularity of the image and calculates the similarity of the correct watermark before and after modifications. Experimental results show that the proposed approach is quite effective in authenticating the origin of an image.

High performance scalable image compression with EBCOT

A new image compression algorithm is proposed, based on independent embedded block coding with optimized truncation of the embedded bit-streams (EBCOT). The algorithm exhibits state-of-the-art compression performance while producing a bit-stream with a rich set of features, including resolution and SNR scalability together with a "random access" property. The algorithm has modest complexity and is suitable for applications involving remote browsing of large compressed images. The algorithm lends itself to explicit optimization with respect to MSE as well as more realistic psychovisual metrics, capable of modeling the spatially varying visual masking phenomenon.