# Ultrafast Video Attention Prediction with Coupled Knowledge Distillation

## TL;DR

This paper introduces UVA-Net, a highly efficient video attention prediction model that uses depth-wise convolutions and knowledge distillation to achieve high speed and low memory usage while maintaining competitive accuracy.

## Contribution

The paper presents UVA-Net, a lightweight, ultrafast video attention prediction network that leverages coupled knowledge distillation from complex teacher models.

## Key findings

- Achieves comparable accuracy to 11 state-of-the-art models.
- Runs at 10,106 FPS on GPU and 404 FPS on CPU.
- Uses only 0.68 MB memory footprint.

## Abstract

Large convolutional neural network models have recently demonstrated impressive performance on video attention prediction. Conventionally, these models are with intensive computation and large memory. To address these issues, we design an extremely light-weight network with ultrafast speed, named UVA-Net. The network is constructed based on depth-wise convolutions and takes low-resolution images as input. However, this straight-forward acceleration method will decrease performance dramatically. To this end, we propose a coupled knowledge distillation strategy to augment and train the network effectively. With this strategy, the model can further automatically discover and emphasize implicit useful cues contained in the data. Both spatial and temporal knowledge learned by the high-resolution complex teacher networks also can be distilled and transferred into the proposed low-resolution light-weight spatiotemporal network. Experimental results show that the performance of our model is comparable to 11 state-of-the-art models in video attention prediction, while it costs only 0.68 MB memory footprint, runs about 10,106 FPS on GPU and 404 FPS on CPU, which is 206 times faster than previous models.