Adaptive weighted fusion with new spatial and temporal fingerprints for improved video copy detection
作者：
Highlights：
• We propose a new fusion method of video fingerprints for video copy detection.
• We propose two kinds of video fingerprints; one is spatial and the other is temporal.
• Weights for fusion are adaptively decided by temporal strength.
• Temporal strengths are available to filter video fingerprints to reduce matching time.
• We provide results comparing the state-of-the arts for video copy detection.
摘要
•We propose a new fusion method of video fingerprints for video copy detection.•We propose two kinds of video fingerprints; one is spatial and the other is temporal.•Weights for fusion are adaptively decided by temporal strength.•Temporal strengths are available to filter video fingerprints to reduce matching time.•We provide results comparing the state-of-the arts for video copy detection.
论文关键词：Video copy detection,Video sequence matching,Modality fusion,Video fingerprint,Weighted adaptive fusion,Spatial and temporal information
论文评审过程：Received 24 January 2013, Revised 14 April 2014, Accepted 2 May 2014, Available online 23 May 2014.
论文官网地址：https://doi.org/10.1016/j.image.2014.05.002
