Unsupervised Image-to-Video Adaptation via Category-aware Flow Memory Bank and Realistic Video Generation
Lists
Only you see this
Abstract
Image-to-Video adaptation is proposed to train a model using labeled images and unlabeled videos to facilitate the classification of unlabeled videos. The latest work synthesizes videos using still images to mitigate the modality gap between images and videos. However, the synthesized videos are not realistic due to the camera movements are only simulated in 2D space. Therefore, we generate realistic videos by simulating arbitrary camera movements in 3D scenes, and then the model can be trained using the generated source videos. Unfortunately, the optical flows from the generated videos have unexpected negative impacts, resulting in suboptimal performance. To address this issue, we propose the Category-aware Flow Memory Bank, which replaces optical flows in source videos with real target flows, and the new composed videos are beneficial for training. In addition, we leverage the video pace prediction task to enhance the speed awareness of the model in order to solve the problem that the model performs poorly in handling some categories with similar appearances but significant speed differences. Our method achieves state-of-the-art performance and comparable performance on three Image-to-Video benchmarks.
BibTeX
@inproceedings{huang2024unsupervised,
author = {Kenan Huang and Junbao Zhuo and Shuhui Wang and Chi Su and Qingming Huang and Huimin Ma},
title = {{Unsupervised Image-to-Video Adaptation via Category-aware Flow Memory Bank and Realistic Video Generation}},
booktitle = {ACM Multimedia},
year = {2024},
doi = {10.1145/3664647.3681063},
}Loading related content...