Search

Copyright of Chinese Journal of Liquid Crystal & Displays is the property of Chinese Journal of Liquid Crystal & Displays and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

Copyright of Journal of Computer Engineering & Applications is the property of Beijing Journal of Computer Engineering & Applications Journal Co Ltd. and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

Copyright of Telecommunication Engineering is the property of Telecommunication Engineering and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

In order to cope with the problem that reversible data hiding in encrypted images(RDHEI) has lower embedding capacity, this paper proposed a high-capacity RDHEI method based on prediction error dual coding. Firstly, to reserve space for secret data, the image owner preprocessed the image using prediction error-based Huffman coding and extended run-length coding, and then encrypted the image. The data hider embedded secret data in the encrypted image. The receiver could extract the secret data accurately according to the data hiding key and recover the image losslessly according to the decryption key, it separated data extraction and image decryption. The experimental results demonstrate that the embedding capacity is effectively improved by prediction error dual coding. [ABSTRACT FROM AUTHOR]

Copyright of Chinese Journal of Liquid Crystal & Displays is the property of Chinese Journal of Liquid Crystal & Displays and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

Aiming at the problems of low embedding rate and insufficient security in the existing reversible information hiding algorithms in the ciphertext domain, this paper proposed a large-capacity reversible information hiding scheme in the ciphertext domain using the correlation between image pixels. Firstly, it used the correlation between image bit planes to reduce redundancy, then used the Peano curve to scan the bit planes, used run-length Huffman coding to compress each bit plane, and then used the high-level information of the image to fill the low-bit space, and finally used the padding message as a hidden key to XOR the secret information to realize embedding. Experimental results show that the method can restore the original image completely reversibly, and the average maximum embedding capacity can reach 2.53 bpp. [ABSTRACT FROM AUTHOR]

To solve the problems of detection process with low detection accuracy and redundancy in the traditional detection method of Marine plankton by artificial features extraction, a multi module fusion single shot detector ( MMFSSD) was proposed based on deep learning technology, The feature information enhancement module was proposed to add the receptive field of network without increasing the network complexity, and the down-sampled image was infused into the module to enhance the low-level feature information of feature graph, The selective feature fusion module was further proposed to learn the weight of fusion in the network and selectively fuse features of different scales, The results of verification test show that the mean average precision values are 80, 70% and 32, 20% on PASCAL VOC and MS COCO test-set, respectively, The mean average precision on PMID2019 data-set reaches 90, 41 %. [ABSTRACT FROM AUTHOR]

To further explore the characteristics of the steganographic noise signal in the adaptive JPEG steganographic images, this paper proposed a noise-aware residual network for JPEG image steganalysis. The proposed network included three components, noise-aware component, noise analysis component and judgement component. First, noise-aware component extracted the noise of an image, it used denoising network to offer more comprehensive information related to steganography. Then, noise analysis component got discriminative features from image noise related information. Last,judgement network gave the final detection result. Moreover, residual connections integrated high-level features and low-level features and prevented gradient vanishing and exploding. Experiments under various conditions show that the network has an advanced performance, and has better generalization ability. [ABSTRACT FROM AUTHOR]

To address the problem of satellite images in the transmission and storage, this paper designed a two-level lossless compress ion algorithm based on sparse representation for satellite images. It replaced the transmission of satellite images by that of sparse coefficients which was created by sparse representation, realizing the first-level compression. Firstly, it pre-processed the non-zero sparse coefficients to realize clustering, and sorted the locations of the original non-zero sparse coefficients by the clustering index. Then, utilizing the result of clustering, it divided the reordered sparse coefficients and the position data into blocks. Finally, it proposed an improved adaptive Huffman coding algorithm to code the blocks of sparse coefficients, while the blocks of their locations were via difference coding followed by improved Huffman coding, and the two-level compression of the image data was accordingly done. Experimental results show that the proposed algorithm is superior to the traditional algorithm, and the compression ratio of the improved algorithm is about 1/3 - 112 times that of the traditional algorithm, which can achieve high lossless compression and high resolution reconstruction of satellite images. [ABSTRACT FROM AUTHOR]

To solve the problems of hierarchical Softmax algorithm in the training process of word vectors with inability of incremental training and inefficient training of massive data, the dynamic hierarchical Softmax algorithm was proposed. By the incremental loading of data samples, an adaptive Huffman coding tree was dynamically constructed by the node adjustment replacement method. To avoid the oscillatory decline of loss function due to the small sample size, the first-order and the second-order moment estimations of the gradient were used to dynamically adjust the parameters update direction and learning rate. The weight variation range and the convergence training network error were reduced by the gradient descent algorithm to improve the training efficiency of the word vector from massive data. The Wikipedia Chinese corpus was adopted as the data to test the training efficiency and quality. The experimental results show that the dynamic hierarchical Softmax algorithm can significantly improve the training efficiency and ensure the quality of word vector training. When the incremental samples are from 10 kB to 1 MB, the training speed is increased about 30 times, which can effectively shorten the training period. [ABSTRACT FROM AUTHOR]

Copyright of Transactions of Beijing Institute of Technology is the property of Beijing University of Technology and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)