NeXtVLAD: An Efficient Neural Network to Aggregate Frame-level Features for Large-scale Video Classification
This work addresses the challenge of efficient video classification for large-scale applications, representing an incremental improvement over existing methods.
The paper tackles the problem of aggregating frame-level features for large-scale video classification by introducing NeXtVLAD, an efficient neural network architecture that decomposes high-dimensional features before applying NetVLAD aggregation, achieving a GAP score of 0.87846 with a single model and 0.88722 with a mixture of three models in the 2nd YouTube-8M challenge.
This paper introduces a fast and efficient network architecture, NeXtVLAD, to aggregate frame-level features into a compact feature vector for large-scale video classification. Briefly speaking, the basic idea is to decompose a high-dimensional feature into a group of relatively low-dimensional vectors with attention before applying NetVLAD aggregation over time. This NeXtVLAD approach turns out to be both effective and parameter efficient in aggregating temporal information. In the 2nd Youtube-8M video understanding challenge, a single NeXtVLAD model with less than 80M parameters achieves a GAP score of 0.87846 in private leaderboard. A mixture of 3 NeXtVLAD models results in 0.88722, which is ranked 3rd over 394 teams. The code is publicly available at https://github.com/linrongc/youtube-8m.