A fast and efficient saliency detection model in video compressed-domain for human fixations prediction

被引:11
作者
Li, Yongjun [1 ,2 ,3 ]
Li, Yunsong [1 ,2 ]
机构
[1] Xidian Univ, Sch Telecommun Engn, State Key Lab Integrated Serv Networks, 2 South Taibai St, Xian 710071, Peoples R China
[2] Xidian Univ, Sch Telecommun Engn, Joint Lab High Speed Multisource Image Coding & P, 2 South Taibai St, Xian 710071, Peoples R China
[3] Henan Univ, Sch Phys & Elect, 1 Jinming St, Kaifeng 475004, Henan, Peoples R China
关键词
Compressed domain; Human fixations detection; Visual saliency; BOTTOM-UP; VISUAL-ATTENTION; TOP-DOWN; VISION; SEARCH;
D O I
10.1007/s11042-016-4118-3
中图分类号
TP [自动化技术、计算机技术];
学科分类号
0812 ;
摘要
Research and application of human fixations detection in video compressed-domain have gained an increasing attention in the latest years. However, both prediction accuracy and computational complexity still remain a challenge. This paper addresses the problem of compressed-domain video human fixations prediction based on saliency detection, and presents a fast and efficient algorithm based on Residual DCT Coefficients Norm (RDCN feature) and Operational Block Description Length (OBDL feature). These two features are directly extracted from the compressed bit-stream with partial decoding, and are normalized. After spatial and temporal filtering, the normalized salient maps are fused by the dynamic fusion coefficients with variation of quantization parameters. Then the fused salient map is worked by Gaussian model whose center is determined by the feature values. The proposed saliency detection model for human fixations prediction combines the accuracy of the pixel-domain saliency detections with the computational efficiency of their compressed-domain counterparts. The validation and comparison are made by several accuracy metrics on two ground truth datasets. Experimental results show that the proposed saliency detection model for human fixations prediction obtains superior performances over several state-of-the-art compressed-domain and pixel-domain algorithms on evaluation metrics. Computationally, our algorithm achieves a speed-up of over 10 times as compared to similar algorithms, which illustrates it appropriate for in-camera saliency estimation.
引用
收藏
页码:26273 / 26295
页数:23
相关论文
共 50 条
  • [21] A Compressed-domain Video Encryption Algorithm for H.264/AVC
    Zhang, P. M.
    Zhu, W. H.
    Kang, Z. M.
    Shi, Z.
    Wang, K. Y.
    INTERNATIONAL CONFERENCE ON COMPUTER SCIENCE AND ENVIRONMENTAL ENGINEERING (CSEE 2015), 2015, : 1377 - 1383
  • [22] SCENE-AWARE SOCCER VIDEO QOE ASSESSMENT - A COMPRESSED-DOMAIN APPROACH
    Li, Fan
    Mei, Yixin
    Liu, Ziyi
    Cosman, Pamela
    2018 IEEE INTERNATIONAL CONFERENCE ON MULTIMEDIA AND EXPO (ICME), 2018,
  • [23] Saliency Detection in the Compressed Domain for Adaptive Image Retargeting
    Fang, Yuming
    Chen, Zhenzhong
    Lin, Weisi
    Lin, Chia-Wen
    IEEE TRANSACTIONS ON IMAGE PROCESSING, 2012, 21 (09) : 3888 - 3901
  • [24] A novel RS-based key frame representation for video mining in Compressed-Domain
    Li Xiang-wei
    Zhang Ming-xin
    Li Xiang-wei
    Zhu Ya-lin
    Xin jin-hong
    WKDD: 2009 SECOND INTERNATIONAL WORKSHOP ON KNOWLEDGE DISCOVERY AND DATA MINING, PROCEEDINGS, 2009, : 199 - +
  • [25] Fast Object Detection in Compressed Video
    Wang, Shiyao
    Lu, Hongchao
    Deng, Zhidong
    2019 IEEE/CVF INTERNATIONAL CONFERENCE ON COMPUTER VISION (ICCV 2019), 2019, : 7103 - 7112
  • [26] Surveillance video synopsis in the compressed domain for fast video browsing
    Wang, Shi-zheng
    Wang, Zhong-yuan
    Hu, Rui-min
    JOURNAL OF VISUAL COMMUNICATION AND IMAGE REPRESENTATION, 2013, 24 (08) : 1431 - 1442
  • [27] An efficient compressed domain video indexing method
    Farahnaz Akrami
    Farzad Zargari
    Multimedia Tools and Applications, 2014, 72 : 705 - 721
  • [28] An efficient compressed domain video indexing method
    Akrami, Farahnaz
    Zargari, Farzad
    MULTIMEDIA TOOLS AND APPLICATIONS, 2014, 72 (01) : 705 - 721
  • [29] Visual Saliency Detection Based on Mutual Information in Compressed Domain
    Gao, Ran
    Tu, Qin
    Xu, Jun
    Lu, Yanping
    Xie, Wei
    Men, Aidong
    2015 VISUAL COMMUNICATIONS AND IMAGE PROCESSING (VCIP), 2015,
  • [30] An efficient approach to extract moving objects by the H.264 compressed-domain features
    Wang, Fu-Ping
    Chung, Wei-Ho
    Kuo, Sy-Yen
    2012 12TH INTERNATIONAL CONFERENCE ON ITS TELECOMMUNICATIONS (ITST-2012), 2012, : 446 - 450