To utilize residual redundancy to reduce the error induced by fading channels and decrease the complexity of the field model to describe the probability structure for residual redundancy, a simplified statistical mode...To utilize residual redundancy to reduce the error induced by fading channels and decrease the complexity of the field model to describe the probability structure for residual redundancy, a simplified statistical model for residual redundancy and a low complexity joint source-channel decoding（JSCD） algorithm are proposed. The complicated residual redundancy in wavelet compressed images is decomposed into several independent 1-D probability check equations composed of Markov chains and it is regarded as a natural channel code with a structure similar to the low density parity check （LDPC） code. A parallel sum-product （SP） and iterative JSCD algorithm is proposed. Simulation results show that the proposed JSCD algorithm can make full use of residual redundancy in different directions to correct errors and improve the peak signal noise ratio （PSNR） of the reconstructed image and reduce the complexity and delay of JSCD. The performance of JSCD is more robust than the traditional separated encoding system with arithmetic coding in the same data rate.展开更多
The better compression rate can be achieved by the traditional vector quantization （VQ） method, and the quality of the recovered image can also be accepted. But the decompressed image quality can not be promoted eff...The better compression rate can be achieved by the traditional vector quantization （VQ） method, and the quality of the recovered image can also be accepted. But the decompressed image quality can not be promoted efficiently, so how to balance the image compression rate and image recovering quality is an important issue, in this paper, an image is transformed by discrete wavelet transform （DWT） to generate its DWT transformed image which can be compressed by the VQ method further. Besides, we compute the values between the DWT transformed image and decompressed DWT transformed image as the difference matrix which is the adjustable basis of the decompressed image quality. By controlling the deviation of the difference matrix, there can be nearly Iossless compression for the VQ method. Experimental results show that when the number of compressed bits by our method is equal to the number of those bits compressed by the VQ method, the quality of our recovered image is better. Moreover, the proposed method has more compression capability comparing with the VQ scheme.展开更多
Due to coarse quantization, block-based discrete cosine transform(BDCT) compression methods usually suffer from visible blocking artifacts at the block boundaries. A novel efficient de-blocking method in DCT domain is...Due to coarse quantization, block-based discrete cosine transform(BDCT) compression methods usually suffer from visible blocking artifacts at the block boundaries. A novel efficient de-blocking method in DCT domain is proposed. A specific criterion for edge detection is given, one-dimensional DCT is applied on each row of the adjacent blocks and the shifted block in smooth region, and the transform coefficients of the shifted block are modified by weighting the average of three coefficients of the block. Mean square difference of slope criterion is used to judge the efficiency of the proposed algorithm. Simulation results show that the new method not only obtains satisfactory image quality, but also maintains high frequency information.展开更多
With the advent of the information security era,it is necessary to guarantee the privacy,accuracy,and dependable transfer of pictures.This study presents a new approach to the encryption and compression of color image...With the advent of the information security era,it is necessary to guarantee the privacy,accuracy,and dependable transfer of pictures.This study presents a new approach to the encryption and compression of color images.It is predicated on 2D compressed sensing(CS)and the hyperchaotic system.First,an optimized Arnold scrambling algorithm is applied to the initial color images to ensure strong security.Then,the processed images are con-currently encrypted and compressed using 2D CS.Among them,chaotic sequences replace traditional random measurement matrices to increase the system’s security.Third,the processed images are re-encrypted using a combination of permutation and diffusion algorithms.In addition,the 2D projected gradient with an embedding decryption(2DPG-ED)algorithm is used to reconstruct images.Compared with the traditional reconstruction algorithm,the 2DPG-ED algorithm can improve security and reduce computational complexity.Furthermore,it has better robustness.The experimental outcome and the performance analysis indicate that this algorithm can withstand malicious attacks and prove the method is effective.展开更多
In the process of image transmission, the famous JPEG and JPEG-2000 compression methods need more transmission time as it is difficult for them to compress the image with a low compression rate. Recently the compresse...In the process of image transmission, the famous JPEG and JPEG-2000 compression methods need more transmission time as it is difficult for them to compress the image with a low compression rate. Recently the compressed sensing（CS） theory was proposed, which has earned great concern as it can compress an image with a low compression rate, meanwhile the original image can be perfectly reconstructed from only a few compressed data. The CS theory is used to transmit the high resolution astronomical image and build the simulation environment where there is communication between the satellite and the Earth. Number experimental results show that the CS theory can effectively reduce the image transmission and reconstruction time. Even with a very low compression rate, it still can recover a higher quality astronomical image than JPEG and JPEG-2000 compression methods.展开更多
In this paper, we are proposing a compression-based multiple color target detection for practical near real-time optical pattern recognition applications. By reducing the size of the color images to its utmost compres...In this paper, we are proposing a compression-based multiple color target detection for practical near real-time optical pattern recognition applications. By reducing the size of the color images to its utmost compression, the speed and the storage of the system are greatly increased. We have used the powerful Fringe-adjusted joint transform correlation technique to successfully detect compression-based multiple targets in colored images. The colored image is decomposed into three fundamental color components images (Red, Green, Blue) and they are separately processed by three-channel correlators. The outputs of the three channels are then combined into a single correlation output. To eliminate the false alarms and zero-order terms due to multiple desired and undesired targets in a scene, we have used the reference shifted phase-encoded and the reference phase-encoded techniques. The performance of the proposed compression-based technique is assessed through many computer simulation tests for images polluted by strong additive Gaussian and Salt & Pepper noises as well as reference occluded images. The robustness of the scheme is demonstrated for severely compressed images (up to 94% ratio), strong noise densities (up to 0.5), and large reference occlusion images (up to 75%).展开更多
Single-shot ultrafast compressed imaging(UCI)is an effective tool for studying ultrafast dynamics in physics,chemistry,or material science because of its excellent high frame rate and large frame number.However,the ra...Single-shot ultrafast compressed imaging(UCI)is an effective tool for studying ultrafast dynamics in physics,chemistry,or material science because of its excellent high frame rate and large frame number.However,the random code(Rcode)used in traditional UCI will lead to low-frequency noise covering high-frequency information due to its uneven sampling interval,which is a great challenge in the fidelity of large-frame reconstruction.Here,a high-frequency enhanced compressed active photography(H-CAP)is proposed.By uniformizing the sampling interval of R-code,H-CAP capture the ultrafast process with a random uniform sampling mode.This sampling mode makes the high-frequency sampling energy dominant,which greatly suppresses the low-frequency noise blurring caused by R-code and achieves high-frequency information of image enhanced.The superior dynamic performance and large-frame reconstruction ability of H-CAP are verified by imaging optical self-focusing effect and static object,respectively.We applied H-CAP to the spatial-temporal characterization of double-pulse induced silicon surface ablation dynamics,which is performed within 220 frames in a single-shot of 300 ps.H-CAP provides a high-fidelity imaging method for observing ultrafast unrepeatable dynamic processes with large frames.展开更多
With the rapid development of digital communication and the widespread use of the Internet of Things,multi-view image compression has attracted increasing attention as a fundamental technology for image data communica...With the rapid development of digital communication and the widespread use of the Internet of Things,multi-view image compression has attracted increasing attention as a fundamental technology for image data communication.Multi-view image compression aims to improve compression efficiency by leveraging correlations between images.However,the requirement of synchronization and inter-image communication at the encoder side poses significant challenges,especially for constrained devices.In this study,we introduce a novel distributed image compression model based on the attention mechanism to address the challenges associated with the availability of side information only during decoding.Our model integrates an encoder network,a quantization module,and a decoder network,to ensure both high compression performance and high-quality image reconstruction.The encoder uses a deep Convolutional Neural Network(CNN)to extract high-level features from the input image,which then pass through the quantization module for further compression before undergoing lossless entropy coding.The decoder of our model consists of three main components that allow us to fully exploit the information within and between images on the decoder side.Specifically,we first introduce a channel-spatial attention module to capture and refine information within individual image feature maps.Second,we employ a semi-coupled convolution module to extract both shared and specific information in images.Finally,a cross-attention module is employed to fuse mutual information extracted from side information.The effectiveness of our model is validated on various datasets,including KITTI Stereo and Cityscapes.The results highlight the superior compression capabilities of our method,surpassing state-of-the-art techniques.展开更多
Nonlinear transforms have significantly advanced learned image compression(LIC),particularly using residual blocks.This transform enhances the nonlinear expression ability and obtain compact feature representation by ...Nonlinear transforms have significantly advanced learned image compression(LIC),particularly using residual blocks.This transform enhances the nonlinear expression ability and obtain compact feature representation by enlarging the receptive field,which indicates how the convolution process extracts features in a high dimensional feature space.However,its functionality is restricted to the spatial dimension and network depth,limiting further improvements in network performance due to insufficient information interaction and representation.Crucially,the potential of high dimensional feature space in the channel dimension and the exploration of network width/resolution remain largely untapped.In this paper,we consider nonlinear transforms from the perspective of feature space,defining high-dimensional feature spaces in different dimensions and investigating the specific effects.Firstly,we introduce the dimension increasing and decreasing transforms in both channel and spatial dimensions to obtain high dimensional feature space and achieve better feature extraction.Secondly,we design a channel-spatial fusion residual transform(CSR),which incorporates multi-dimensional transforms for a more effective representation.Furthermore,we simplify the proposed fusion transform to obtain a slim architecture(CSR-sm),balancing network complexity and compression performance.Finally,we build the overall network with stacked CSR transforms to achieve better compression and reconstruction.Experimental results demonstrate that the proposed method can achieve superior ratedistortion performance compared to the existing LIC methods and traditional codecs.Specifically,our proposed method achieves 9.38%BD-rate reduction over VVC on Kodak dataset.展开更多
A blind digital image forensic method for detecting copy-paste forgery between JPEG images was proposed.Two copy-paste tampering scenarios were introduced at first:the tampered image was saved in an uncompressed forma...A blind digital image forensic method for detecting copy-paste forgery between JPEG images was proposed.Two copy-paste tampering scenarios were introduced at first:the tampered image was saved in an uncompressed format or in a JPEG compressed format.Then the proposed detection method was analyzed and simulated for all the cases of the two tampering scenarios.The tampered region is detected by computing the averaged sum of absolute difference(ASAD) images between the examined image and a resaved JPEG compressed image at different quality factors.The experimental results show the advantages of the proposed method:capability of detecting small and/or multiple tampered regions,simple computation,and hence fast speed in processing.展开更多
The isotherm is an important feature of infrared satellite cloud images (ISCI), which can directly reveal substantial information of cloud systems. The isotherm extraction of ISCI can remove the redundant information ...The isotherm is an important feature of infrared satellite cloud images (ISCI), which can directly reveal substantial information of cloud systems. The isotherm extraction of ISCI can remove the redundant information and therefore helps to compress the information of ISCI. In this paper, an isotherm extraction method is presented. The main aggregate of clouds can be segmented based on mathematical morphology. T algorithm and IP algorithm are then applied to extract the isotherms from the main aggregate of clouds. A concrete example for the extraction of isotherm based on IBM SP2 is described. The result shows that this is a high efficient algorithm. It can be used in feature extractions of infrared images for weather forecasts.展开更多
[Objective] The aim was to present a proposal about a new image compression technology, in order to make the image be able to be stored in a smaller space and be transmitted with smaller bit rate on the premise of gua...[Objective] The aim was to present a proposal about a new image compression technology, in order to make the image be able to be stored in a smaller space and be transmitted with smaller bit rate on the premise of guaranteeing image quality in the rape crop monitoring system in Qinling Mountains. [Method] In the proposal, the color image was divided into brightness images with three fundamental colors, followed by sub-image division and DCT treatment. Then, coefficients of transform domain were quantized, and encoded and compressed as per Huffman coding. Finally, decompression was conducted through inverse process and decompressed images were matched. [Result] The simulation results show that when compression ratio of the color image of rape crops was 11.972 3∶1, human can not distinguish the differences between the decompressed images and the source images with naked eyes; when ratio was as high as 53.565 6∶1, PSNR was still above 30 dD,encoding efficiency achieved over 0.78 and redundancy was less than 0.22. [Conclusion] The results indicate that the proposed color image compression technology can achieve higher compression ratio on the premise of good image quality. In addition, image encoding quality and decompressed images achieved better results, which fully met requirement of image storage and transmission in monitoring system of rape crop in the Qinling Mountains.展开更多
By investigating the limitation of existing wavelet tree based image compression methods, we propose a novel wavelet fractal image compression method in this paper. Briefly, the initial errors are appointed given the ...By investigating the limitation of existing wavelet tree based image compression methods, we propose a novel wavelet fractal image compression method in this paper. Briefly, the initial errors are appointed given the different levels of importance accorded the frequency sublevel band wavelet coefficients. Higher frequency sublevel bands would lead to larger initial errors. As a result, the sizes of sublevel blocks and super blocks would be changed according to the initial errors. The matching sizes between sublevel blocks and super blocks would be changed according to the permitted errors and compression rates. Systematic analyses are performed and the experimental results demonstrate that the proposed method provides a satisfactory performance with a clearly increasing rate of compression and speed of encoding without reducing SNR and the quality of decoded images. Simulation results show that our method is superior to the traditional wavelet tree based methods of fractal image compression.展开更多
Many classical encoding algorithms of vector quantization （VQ） of image compression that can obtain global optimal solution have computational complexity O（N）. A pure quantum VQ encoding algorithm with probability...Many classical encoding algorithms of vector quantization （VQ） of image compression that can obtain global optimal solution have computational complexity O（N）. A pure quantum VQ encoding algorithm with probability of success near 100% has been proposed, that performs operations 45√N times approximately. In this paper, a hybrid quantum VQ encoding algorithm between the classical method and the quantum algorithm is presented. The number of its operations is less than √N for most images, and it is more efficient than the pure quantum algorithm.展开更多
Based on Jacquin's work. this paper presents an adaptive block-based fractal image coding scheme. Firstly. masking functions are used to classify range blocks and weight the mean Square error (MSE) of images. Seco...Based on Jacquin's work. this paper presents an adaptive block-based fractal image coding scheme. Firstly. masking functions are used to classify range blocks and weight the mean Square error (MSE) of images. Secondly, an adaptive block partition scheme is introduced by developing the quadtree partition method. Thirdly. a piecewise uniform quantization strategy is appled to quantize the luminance shifting. Finally. experiment results are shown and compared with what reported by Jacquin and Lu to verify the validity of the methods addressed by the authors.展开更多
This paper utilizes a spatial texture correlation and the intelligent classification algorithm （ICA） search strategy to speed up the encoding process and improve the bit rate for fractal image compression. Texture f...This paper utilizes a spatial texture correlation and the intelligent classification algorithm （ICA） search strategy to speed up the encoding process and improve the bit rate for fractal image compression. Texture features is one of the most important properties for the representation of an image. Entropy and maximum entry from co-occurrence matrices are used for representing texture features in an image. For a range block, concerned domain blocks of neighbouring range blocks with similar texture features can be searched. In addition, domain blocks with similar texture features are searched in the ICA search process. Experiments show that in comparison with some typical methods, the proposed algorithm significantly speeds up the encoding process and achieves a higher compression ratio, with a slight diminution in the quality of the reconstructed image; in comparison with a spatial correlation scheme, the proposed scheme spends much less encoding time while the compression ratio and the quality of the reconstructed image are almost the same.展开更多
A chaos-based cryptosystem for fractal image coding is proposed. The Renyi chaotic map is employed to determine the order of processing the range blocks and to generate the keystream for masking the encoded sequence. ...A chaos-based cryptosystem for fractal image coding is proposed. The Renyi chaotic map is employed to determine the order of processing the range blocks and to generate the keystream for masking the encoded sequence. Compared with the standard approach of fraetal image coding followed by the Advanced Encryption Standard, our scheme offers a higher sensitivity to both plaintext and ciphertext at a comparable operating efficiency. The keystream generated by the Renyi chaotic map passes the randomness tests set by the United States National Institute of Standards and Technology, and so the proposed scheme is sensitive to the key.展开更多
A new method using plane fitting to decide whether a domain block is similar enough to a given range block is proposed in this paper. First, three coefficients are computed for describing each range and domain block. ...A new method using plane fitting to decide whether a domain block is similar enough to a given range block is proposed in this paper. First, three coefficients are computed for describing each range and domain block. Then, the best-matched one for every range block is obtained by analysing the relation between their coefficients. Experimental results show that the proposed method can shorten encoding time markedly, while the retrieved image quality is still acceptable. In the decoding step, a kind of simple line fitting on block boundaries is used to reduce blocking effects. At the same time, the proposed method can also achieve a high compression ratio.展开更多
A nonlinear data analysis algorithm, namely empirical data decomposition （EDD） is proposed, which can perform adaptive analysis of observed data. Analysis filter, which is not a linear constant coefficient filter, i...A nonlinear data analysis algorithm, namely empirical data decomposition （EDD） is proposed, which can perform adaptive analysis of observed data. Analysis filter, which is not a linear constant coefficient filter, is automatically determined by observed data, and is able to implement multi-resolution analysis as wavelet transform. The algorithm is suitable for analyzing non-stationary data and can effectively wipe off the relevance of observed data. Then through discussing the applications of EDD in image compression, the paper presents a 2-dimension data decomposition framework and makes some modifications of contexts used by Embedded Block Coding with Optimized Truncation （EBCOT） . Simulation results show that EDD is more suitable for non-stationary image data compression.展开更多
The paper presents a class of nonlinear adaptive wavelet transforms for lossless image compression. In update step of the lifting the different operators are chosen by the local gradient of original image. A nonlinear...The paper presents a class of nonlinear adaptive wavelet transforms for lossless image compression. In update step of the lifting the different operators are chosen by the local gradient of original image. A nonlinear morphological predictor follows the update adaptive lifting to result in fewer large wavelet coefficients near edges for reducing coding. The nonlinear adaptive wavelet transforms can also allow perfect reconstruction without any overhead cost. Experiment results are given to show lower entropy of the adaptive transformed images than those of the non-adaptive case and great applicable potentiality in lossless image compresslon.展开更多
文摘To utilize residual redundancy to reduce the error induced by fading channels and decrease the complexity of the field model to describe the probability structure for residual redundancy, a simplified statistical model for residual redundancy and a low complexity joint source-channel decoding（JSCD） algorithm are proposed. The complicated residual redundancy in wavelet compressed images is decomposed into several independent 1-D probability check equations composed of Markov chains and it is regarded as a natural channel code with a structure similar to the low density parity check （LDPC） code. A parallel sum-product （SP） and iterative JSCD algorithm is proposed. Simulation results show that the proposed JSCD algorithm can make full use of residual redundancy in different directions to correct errors and improve the peak signal noise ratio （PSNR） of the reconstructed image and reduce the complexity and delay of JSCD. The performance of JSCD is more robust than the traditional separated encoding system with arithmetic coding in the same data rate.
文摘The better compression rate can be achieved by the traditional vector quantization （VQ） method, and the quality of the recovered image can also be accepted. But the decompressed image quality can not be promoted efficiently, so how to balance the image compression rate and image recovering quality is an important issue, in this paper, an image is transformed by discrete wavelet transform （DWT） to generate its DWT transformed image which can be compressed by the VQ method further. Besides, we compute the values between the DWT transformed image and decompressed DWT transformed image as the difference matrix which is the adjustable basis of the decompressed image quality. By controlling the deviation of the difference matrix, there can be nearly Iossless compression for the VQ method. Experimental results show that when the number of compressed bits by our method is equal to the number of those bits compressed by the VQ method, the quality of our recovered image is better. Moreover, the proposed method has more compression capability comparing with the VQ scheme.
基金Science and Technology Project of Guangdong Province(2006A10201003) 2005 Jinan University StartupProject(51205067) Soft Science Project of Guangdong Province(2006B70103011)
文摘Due to coarse quantization, block-based discrete cosine transform(BDCT) compression methods usually suffer from visible blocking artifacts at the block boundaries. A novel efficient de-blocking method in DCT domain is proposed. A specific criterion for edge detection is given, one-dimensional DCT is applied on each row of the adjacent blocks and the shifted block in smooth region, and the transform coefficients of the shifted block are modified by weighting the average of three coefficients of the block. Mean square difference of slope criterion is used to judge the efficiency of the proposed algorithm. Simulation results show that the new method not only obtains satisfactory image quality, but also maintains high frequency information.
基金This work was supported in part by the National Natural Science Foundation of China under Grants 71571091,71771112the State Key Laboratory of Synthetical Automation for Process Industries Fundamental Research Funds under Grant PAL-N201801the Excellent Talent Training Project of University of Science and Technology Liaoning under Grant 2019RC05.
文摘With the advent of the information security era,it is necessary to guarantee the privacy,accuracy,and dependable transfer of pictures.This study presents a new approach to the encryption and compression of color images.It is predicated on 2D compressed sensing(CS)and the hyperchaotic system.First,an optimized Arnold scrambling algorithm is applied to the initial color images to ensure strong security.Then,the processed images are con-currently encrypted and compressed using 2D CS.Among them,chaotic sequences replace traditional random measurement matrices to increase the system’s security.Third,the processed images are re-encrypted using a combination of permutation and diffusion algorithms.In addition,the 2D projected gradient with an embedding decryption(2DPG-ED)algorithm is used to reconstruct images.Compared with the traditional reconstruction algorithm,the 2DPG-ED algorithm can improve security and reduce computational complexity.Furthermore,it has better robustness.The experimental outcome and the performance analysis indicate that this algorithm can withstand malicious attacks and prove the method is effective.
文摘In the process of image transmission, the famous JPEG and JPEG-2000 compression methods need more transmission time as it is difficult for them to compress the image with a low compression rate. Recently the compressed sensing（CS） theory was proposed, which has earned great concern as it can compress an image with a low compression rate, meanwhile the original image can be perfectly reconstructed from only a few compressed data. The CS theory is used to transmit the high resolution astronomical image and build the simulation environment where there is communication between the satellite and the Earth. Number experimental results show that the CS theory can effectively reduce the image transmission and reconstruction time. Even with a very low compression rate, it still can recover a higher quality astronomical image than JPEG and JPEG-2000 compression methods.
文摘In this paper, we are proposing a compression-based multiple color target detection for practical near real-time optical pattern recognition applications. By reducing the size of the color images to its utmost compression, the speed and the storage of the system are greatly increased. We have used the powerful Fringe-adjusted joint transform correlation technique to successfully detect compression-based multiple targets in colored images. The colored image is decomposed into three fundamental color components images (Red, Green, Blue) and they are separately processed by three-channel correlators. The outputs of the three channels are then combined into a single correlation output. To eliminate the false alarms and zero-order terms due to multiple desired and undesired targets in a scene, we have used the reference shifted phase-encoded and the reference phase-encoded techniques. The performance of the proposed compression-based technique is assessed through many computer simulation tests for images polluted by strong additive Gaussian and Salt & Pepper noises as well as reference occluded images. The robustness of the scheme is demonstrated for severely compressed images (up to 94% ratio), strong noise densities (up to 0.5), and large reference occlusion images (up to 75%).
基金supported by the National Science Foundation of China(No.12127806,No.62175195 and No.12304382)the International Joint Research Laboratory for Micro/Nano Manufacturing and Measurement Technologies.
文摘Single-shot ultrafast compressed imaging(UCI)is an effective tool for studying ultrafast dynamics in physics,chemistry,or material science because of its excellent high frame rate and large frame number.However,the random code(Rcode)used in traditional UCI will lead to low-frequency noise covering high-frequency information due to its uneven sampling interval,which is a great challenge in the fidelity of large-frame reconstruction.Here,a high-frequency enhanced compressed active photography(H-CAP)is proposed.By uniformizing the sampling interval of R-code,H-CAP capture the ultrafast process with a random uniform sampling mode.This sampling mode makes the high-frequency sampling energy dominant,which greatly suppresses the low-frequency noise blurring caused by R-code and achieves high-frequency information of image enhanced.The superior dynamic performance and large-frame reconstruction ability of H-CAP are verified by imaging optical self-focusing effect and static object,respectively.We applied H-CAP to the spatial-temporal characterization of double-pulse induced silicon surface ablation dynamics,which is performed within 220 frames in a single-shot of 300 ps.H-CAP provides a high-fidelity imaging method for observing ultrafast unrepeatable dynamic processes with large frames.
基金supported by the National Natural Science Foundation of China(Key Program)(No.11932013)the Tianjin Science and Technology Plan Project(No.22PTZWHZ00040)。
文摘With the rapid development of digital communication and the widespread use of the Internet of Things,multi-view image compression has attracted increasing attention as a fundamental technology for image data communication.Multi-view image compression aims to improve compression efficiency by leveraging correlations between images.However,the requirement of synchronization and inter-image communication at the encoder side poses significant challenges,especially for constrained devices.In this study,we introduce a novel distributed image compression model based on the attention mechanism to address the challenges associated with the availability of side information only during decoding.Our model integrates an encoder network,a quantization module,and a decoder network,to ensure both high compression performance and high-quality image reconstruction.The encoder uses a deep Convolutional Neural Network(CNN)to extract high-level features from the input image,which then pass through the quantization module for further compression before undergoing lossless entropy coding.The decoder of our model consists of three main components that allow us to fully exploit the information within and between images on the decoder side.Specifically,we first introduce a channel-spatial attention module to capture and refine information within individual image feature maps.Second,we employ a semi-coupled convolution module to extract both shared and specific information in images.Finally,a cross-attention module is employed to fuse mutual information extracted from side information.The effectiveness of our model is validated on various datasets,including KITTI Stereo and Cityscapes.The results highlight the superior compression capabilities of our method,surpassing state-of-the-art techniques.
基金supported by the Key Program of the National Natural Science Foundation of China(Grant No.62031013)Guangdong Province Key Construction Discipline Scientific Research Capacity Improvement Project(Grant No.2022ZDJS117).
文摘Nonlinear transforms have significantly advanced learned image compression(LIC),particularly using residual blocks.This transform enhances the nonlinear expression ability and obtain compact feature representation by enlarging the receptive field,which indicates how the convolution process extracts features in a high dimensional feature space.However,its functionality is restricted to the spatial dimension and network depth,limiting further improvements in network performance due to insufficient information interaction and representation.Crucially,the potential of high dimensional feature space in the channel dimension and the exploration of network width/resolution remain largely untapped.In this paper,we consider nonlinear transforms from the perspective of feature space,defining high-dimensional feature spaces in different dimensions and investigating the specific effects.Firstly,we introduce the dimension increasing and decreasing transforms in both channel and spatial dimensions to obtain high dimensional feature space and achieve better feature extraction.Secondly,we design a channel-spatial fusion residual transform(CSR),which incorporates multi-dimensional transforms for a more effective representation.Furthermore,we simplify the proposed fusion transform to obtain a slim architecture(CSR-sm),balancing network complexity and compression performance.Finally,we build the overall network with stacked CSR transforms to achieve better compression and reconstruction.Experimental results demonstrate that the proposed method can achieve superior ratedistortion performance compared to the existing LIC methods and traditional codecs.Specifically,our proposed method achieves 9.38%BD-rate reduction over VVC on Kodak dataset.
基金Project(61172184) supported by the National Natural Science Foundation of ChinaProject(200902482) supported by China Postdoctoral Science Foundation Specially Funded ProjectProject(12JJ6062) supported by the Natural Science Foundation of Hunan Province,China
文摘A blind digital image forensic method for detecting copy-paste forgery between JPEG images was proposed.Two copy-paste tampering scenarios were introduced at first:the tampered image was saved in an uncompressed format or in a JPEG compressed format.Then the proposed detection method was analyzed and simulated for all the cases of the two tampering scenarios.The tampered region is detected by computing the averaged sum of absolute difference(ASAD) images between the examined image and a resaved JPEG compressed image at different quality factors.The experimental results show the advantages of the proposed method:capability of detecting small and/or multiple tampered regions,simple computation,and hence fast speed in processing.
文摘The isotherm is an important feature of infrared satellite cloud images (ISCI), which can directly reveal substantial information of cloud systems. The isotherm extraction of ISCI can remove the redundant information and therefore helps to compress the information of ISCI. In this paper, an isotherm extraction method is presented. The main aggregate of clouds can be segmented based on mathematical morphology. T algorithm and IP algorithm are then applied to extract the isotherms from the main aggregate of clouds. A concrete example for the extraction of isotherm based on IBM SP2 is described. The result shows that this is a high efficient algorithm. It can be used in feature extractions of infrared images for weather forecasts.
基金Supported by Special Fund for Scientific Research of Shannxi Education Department(No:2010JK463)Shaanxi Natural Science Foundation(2011JE012)~~
文摘[Objective] The aim was to present a proposal about a new image compression technology, in order to make the image be able to be stored in a smaller space and be transmitted with smaller bit rate on the premise of guaranteeing image quality in the rape crop monitoring system in Qinling Mountains. [Method] In the proposal, the color image was divided into brightness images with three fundamental colors, followed by sub-image division and DCT treatment. Then, coefficients of transform domain were quantized, and encoded and compressed as per Huffman coding. Finally, decompression was conducted through inverse process and decompressed images were matched. [Result] The simulation results show that when compression ratio of the color image of rape crops was 11.972 3∶1, human can not distinguish the differences between the decompressed images and the source images with naked eyes; when ratio was as high as 53.565 6∶1, PSNR was still above 30 dD,encoding efficiency achieved over 0.78 and redundancy was less than 0.22. [Conclusion] The results indicate that the proposed color image compression technology can achieve higher compression ratio on the premise of good image quality. In addition, image encoding quality and decompressed images achieved better results, which fully met requirement of image storage and transmission in monitoring system of rape crop in the Qinling Mountains.
基金Project 60571049 supported by the National Natural Science Foundation of China
文摘By investigating the limitation of existing wavelet tree based image compression methods, we propose a novel wavelet fractal image compression method in this paper. Briefly, the initial errors are appointed given the different levels of importance accorded the frequency sublevel band wavelet coefficients. Higher frequency sublevel bands would lead to larger initial errors. As a result, the sizes of sublevel blocks and super blocks would be changed according to the initial errors. The matching sizes between sublevel blocks and super blocks would be changed according to the permitted errors and compression rates. Systematic analyses are performed and the experimental results demonstrate that the proposed method provides a satisfactory performance with a clearly increasing rate of compression and speed of encoding without reducing SNR and the quality of decoded images. Simulation results show that our method is superior to the traditional wavelet tree based methods of fractal image compression.
文摘Many classical encoding algorithms of vector quantization （VQ） of image compression that can obtain global optimal solution have computational complexity O（N）. A pure quantum VQ encoding algorithm with probability of success near 100% has been proposed, that performs operations 45√N times approximately. In this paper, a hybrid quantum VQ encoding algorithm between the classical method and the quantum algorithm is presented. The number of its operations is less than √N for most images, and it is more efficient than the pure quantum algorithm.
文摘Based on Jacquin's work. this paper presents an adaptive block-based fractal image coding scheme. Firstly. masking functions are used to classify range blocks and weight the mean Square error (MSE) of images. Secondly, an adaptive block partition scheme is introduced by developing the quadtree partition method. Thirdly. a piecewise uniform quantization strategy is appled to quantize the luminance shifting. Finally. experiment results are shown and compared with what reported by Jacquin and Lu to verify the validity of the methods addressed by the authors.
基金supported by the National Natural Science Foundation of China (Grant Nos. 60573172 and 60973152)the Superior University Doctor Subject Special Scientific Research Foundation of China (Grant No. 20070141014)the Natural Science Foundation of Liaoning Province of China (Grant No. 20082165)
文摘This paper utilizes a spatial texture correlation and the intelligent classification algorithm （ICA） search strategy to speed up the encoding process and improve the bit rate for fractal image compression. Texture features is one of the most important properties for the representation of an image. Entropy and maximum entry from co-occurrence matrices are used for representing texture features in an image. For a range block, concerned domain blocks of neighbouring range blocks with similar texture features can be searched. In addition, domain blocks with similar texture features are searched in the ICA search process. Experiments show that in comparison with some typical methods, the proposed algorithm significantly speeds up the encoding process and achieves a higher compression ratio, with a slight diminution in the quality of the reconstructed image; in comparison with a spatial correlation scheme, the proposed scheme spends much less encoding time while the compression ratio and the quality of the reconstructed image are almost the same.
基金Project supported by the Research Grants Council of the Hong Kong Special Administrative Region,China(Grant No.CityU123009)
文摘A chaos-based cryptosystem for fractal image coding is proposed. The Renyi chaotic map is employed to determine the order of processing the range blocks and to generate the keystream for masking the encoded sequence. Compared with the standard approach of fraetal image coding followed by the Advanced Encryption Standard, our scheme offers a higher sensitivity to both plaintext and ciphertext at a comparable operating efficiency. The keystream generated by the Renyi chaotic map passes the randomness tests set by the United States National Institute of Standards and Technology, and so the proposed scheme is sensitive to the key.
基金Project supported by the National Natural Science Foundation of China (Grant Nos. 61173183, 60973152, and 60573172)the Special Scientific Research Fund for the Doctoral Program of Higher Education of China (Grant No. 20070141014)the Natural Science Foundation of Liaoning Province, China (Grant No. 20082165)
文摘A new method using plane fitting to decide whether a domain block is similar enough to a given range block is proposed in this paper. First, three coefficients are computed for describing each range and domain block. Then, the best-matched one for every range block is obtained by analysing the relation between their coefficients. Experimental results show that the proposed method can shorten encoding time markedly, while the retrieved image quality is still acceptable. In the decoding step, a kind of simple line fitting on block boundaries is used to reduce blocking effects. At the same time, the proposed method can also achieve a high compression ratio.
基金This project was supported by the National Natural Science Foundation of China (60532060)Hainan Education Bureau Research Project (Hjkj200602)Hainan Natural Science Foundation (80551).
文摘A nonlinear data analysis algorithm, namely empirical data decomposition （EDD） is proposed, which can perform adaptive analysis of observed data. Analysis filter, which is not a linear constant coefficient filter, is automatically determined by observed data, and is able to implement multi-resolution analysis as wavelet transform. The algorithm is suitable for analyzing non-stationary data and can effectively wipe off the relevance of observed data. Then through discussing the applications of EDD in image compression, the paper presents a 2-dimension data decomposition framework and makes some modifications of contexts used by Embedded Block Coding with Optimized Truncation （EBCOT） . Simulation results show that EDD is more suitable for non-stationary image data compression.
基金Supported by the National Natural Science Foundation of China (69983005)
文摘The paper presents a class of nonlinear adaptive wavelet transforms for lossless image compression. In update step of the lifting the different operators are chosen by the local gradient of original image. A nonlinear morphological predictor follows the update adaptive lifting to result in fewer large wavelet coefficients near edges for reducing coding. The nonlinear adaptive wavelet transforms can also allow perfect reconstruction without any overhead cost. Experiment results are given to show lower entropy of the adaptive transformed images than those of the non-adaptive case and great applicable potentiality in lossless image compresslon.
