The Computer Vision group at Yahoo Labs, the Internet company’s research arm, has published a data set it used to train the algorithms that allow its video service to automatically interpret and summarize videos. The algorithms, known as TVSum, allow Yahoo’s software to identify the most important scenes in videos and condense videos into short, summarized clips. The data set includes 50 videos with annotations indicating important content to help researchers test their own computer vision and summarization algorithms without the need for extensive user studies.
Image: cveradasouza.
