Audiovisual Highlight Detection in Videos
This work addresses video highlight detection for content creators and viewers, but it is incremental as it builds on existing methods with feature combinations and transfer learning.
The paper tackles highlight detection in videos by combining visual, audiovisual, and audio features, finding that visual features are most informative and audiovisual features improve over visual-only, with results showing knowledge transfer from video summarization to highlight detection.
In this paper, we test the hypothesis that interesting events in unstructured videos are inherently audiovisual. We combine deep image representations for object recognition and scene understanding with representations from an audiovisual affect recognition model. To this set, we include content agnostic audio-visual synchrony representations and mel-frequency cepstral coefficients to capture other intrinsic properties of audio. These features are used in a modular supervised model. We present results from two experiments: efficacy study of single features on the task, and an ablation study where we leave one feature out at a time. For the video summarization task, our results indicate that the visual features carry most information, and including audiovisual features improves over visual-only information. To better study the task of highlight detection, we run a pilot experiment with highlights annotations for a small subset of video clips and fine-tune our best model on it. Results indicate that we can transfer knowledge from the video summarization task to a model trained specifically for the task of highlight detection.