- Our method enables us to automatically detect and segment object-like regions from videos without any manually annotated labels.
- We utilize visual saliency as a prior distribution of region segmentation instead of manually annotated labels.
- A prior distribution for every frame is updated with previous segmentation results, combined with prior information coming from visual saliency.
- We introduce CUDA implementation to accelerate the computation of prior distributions and feature likelihoods, resulting in achieving around 10fps in a mobile PC with CUDA-compatible graphics boards.
Data
This dataset contains 10 videos as inputs, and segmented image sequences as ground-truth.
Required
Any report or publication using this data should cite its use as the following publication:
- Ken Fukuchi, Kouji Miyazato, Akisato Kimura, Shigeru Takagi and Junji Yamato "Saliency-based video segmentation with graph cuts and sequentially updated priors," Proc. International Conference on Multimedia and Expo (ICME2009), pp.638--641, New York, New York, USA, June-July 2009.
Detailed description
Videos : 10 uncompressed AVI clips of natural scenes with 12 fps, including at least one target objects or something others. Length varies 5-10 seconds.
Ground-truth: 10 sets of JPEG images, each corresponds to an input video. Segmented images are provided for almost all the frames excluding first 15 frames.
Download
videoSegmentationData.zip123332.9KB
Example
Publication
- Ken Fukuchi, Kouji Miyazato, Akisato Kimura, Shigeru Takagi and Junji Yamato "Saliency-based video segmentation with graph cuts and sequentially updated priors," Proc. International Conference on Multimedia and Expo (ICME2009), pp.638--641, New York, New York, USA, June-July 2009.