Current Issue : July - September Volume : 2019 Issue Number : 3 Articles : 5 Articles
In the wild, wireless multimedia sensor network (WMSN) communication has limited\nbandwidth and the transmission of wildlife monitoring images always suffers signal interference,\nwhich is time-consuming, or sometimes even causes failure. Generally, only part of each wildlife\nimage is valuable, therefore, if we could transmit the images according to the importance of the\ncontent, the above issues can be avoided. Inspired by the progressive transmission strategy, we\npropose a hierarchical coding progressive transmission method in this paper, which can transmit\nthe saliency object region (i.e. the animal) and its background with different coding strategies and\npriorities. Specifically, we firstly construct a convolution neural network via the MobileNet model for\nthe detection of the saliency object region and obtaining the mask on wildlife. Then, according to\nthe importance of wavelet coefficients, set partitioned in hierarchical tree (SPIHT) lossless coding\nis utilized to transmit the saliency image which ensures the transmission accuracy of the wildlife\nregion. After that, the background region left over is transmitted via the Embedded ZerotreeWavelets\n(EZW) lossy coding strategy, to improve the transmission efficiency. To verify the efficiency of our\nalgorithm, a demonstration of the transmission of field-captured wildlife images is presented. Further,\ncomparison of results with existing EZW and discrete cosine transform (DCT) algorithms shows\nthat the proposed algorithm improves the peak signal to noise ratio (PSNR) and structural similarity \nindex (SSIM) by 21.11%, 14.72% and 9.47%, 6.25%, respectively....
Objective image quality assessment (IQA) is imperative in the current multimedia-intensive\nworld, in order to assess the visual quality of an image at close to a human level of ability.\nMany parameters such as color intensity, structure, sharpness, contrast, presence of an object, etc.,\ndraw human attention to an image. Psychological vision research suggests that human vision is\nbiased to the center area of an image and display screen. As a result, if the center part contains\nany visually salient information, it draws human attention even more and any distortion in that\npart will be better perceived than other parts. To the best of our knowledge, previous IQA methods\nhave not considered this fact. In this paper, we propose a full reference image quality assessment\n(FR-IQA) approach using visual saliency and contrast; however, we give extra attention to the center\nby increasing the sensitivity of the similarity maps in that region. We evaluated our method on three\nlarge-scale popular benchmark databases used by most of the current IQA researchers (TID2008,\nCSIQ and LIVE), having a total of 3345 distorted images with 28 different kinds of distortions.\nOur method is compared with 13 state-of-the-art approaches. This comparison reveals the stronger\ncorrelation of our method with human-evaluated values. The prediction-of-quality score is consistent\nfor distortion specific as well as distortion independent cases. Moreover, faster processing makes it\napplicable to any real-time application....
The ubiquity of data, including multi-media data such as images, enables easy mining and\nanalysis of such data. However, such an analysis might involve the use of sensitive data such as medical\nrecords (including radiological images) and financial records. Privacy-preserving machine learning is\nan approach that is aimed at the analysis of such data in such a way that privacy is not compromised.\nThere are various privacy-preserving data analysis approaches such as k-anonymity, l-diversity, t-closeness\nand Differential Privacy (DP). Currently, DP is a golden standard of privacy-preserving data analysis\ndue to its robustness against background knowledge attacks. In this paper, we report a scheme for\nprivacy-preserving image classification using Support Vector Machine (SVM) and DP. SVM is chosen as\na classification algorithm because unlike variants of artificial neural networks, it converges to a global\noptimum. SVM kernels used are linear and Radial Basis Function (RBF), while................
In recent years, the rapid development of surveillance information in closed-circuit\ntelevision (CCTV) has become an indispensable element in security systems. Several CCTV systems\ndesigned for video compression and encryption need to improve for the best performance and\ndifferent security levels. Specially, the advent of 360 video makes the CCTV promising for surveillance\nwithout any blind areas. Compared to current systems, 360 CCTV requires the large bandwidth with\nlow latency to run smoothly. Therefore, to improve the system performance, it needs to be more robust\nto run smoothly. Video transmission and transcoding is an essential process in converting codecs,\nchanging bitrates or resizing the resolution for 360 videos. High-performance transcoding is one of\nthe key factors of real time CCTV stream. Additionally, the security of video streams from cameras\nto endpoints is also an important priority in CCTV research. In this paper, a real-time transcoding\nsystem designed with the ARIA block cipher encryption algorithm is presented. Experimental\nresults show that the proposed method achieved approximately 200% speedup compared to libx265\nFFmpeg in transcoding task, and it could handle multiple transcoding sessions simultaneously at\nhigh performance for both live 360 CCTV system and existing 2D/3D CCTV system....
With the rapid development of wireless networks, multiple network interfaces are gradually being designed into more and more\nmobile devices. When it comes to data delivery, Stream Control Transmission Protocol (SCTP)-based Concurrent Multipath\nTransfer (CMT) has proven to be quite useful solution for multiple home networks, and it could become the key transport protocol\nfor the next generation of wireless communications.The CMTdelay caused by data rearrangement has been noticed by researchers,\nbut they have seldom considered the frequent occurrence of packet loss that occurs in the high-loss networks. In this paper, we\nproposed an original loss-aware solution for multipath concurrent transmission (CMT-LA) that achieves the following goals: (1)\nidentifying packet loss on all paths, (2) distributing packets adaptively acrossmultiple available paths according to their packet loss\nand loss variation, and (3) maintaining the features of bandwidth aggregation and parallel transmission of CMT while improving\nthe throughput performance. The results of our simulations showed that the proposed CMT-LA reduces reordering delay and\nunnecessary fast retransmissions, thereby demonstrating that CMT-LA is a more efficient data delivery scheme than classic CMT....
Loading....