VideoPrism: A Foundational Visual Encoder for Video Understanding.
Long ZhaoNitesh Bharadwaj GundavarapuLiangzhe YuanHao ZhouShen YanJennifer J. SunLuke FriedmanRui QianTobias WeyandYue ZhaoRachel HornungFlorian SchroffMing-Hsuan YangDavid A. RossHuisheng WangHartwig AdamMikhail SirotenkoTing LiuBoqing GongPublished in: CoRR (2024)
Keyphrases
- visual analysis
- visual cues
- video data
- visual data
- video sequences
- news video
- video search
- video streams
- low complexity
- visual representation
- video encoder
- video database
- multimedia
- video encoding
- bit rate
- mpeg standard
- visual information
- video content
- video compression
- video frames
- real time video
- multimedia data
- temporal correlation
- spatial and temporal
- visual features
- visual concepts
- low level
- channel bandwidth
- real time
- visual saliency
- semantic concepts
- video analysis
- motion compensation
- image quality
- image classification