# Exploring Temporal Information for Improved Video Understanding

## TL;DR

This dissertation explores leveraging temporal information in videos to enhance understanding, proposing novel frameworks for action recognition without optical flow and for semantic segmentation via data augmentation, leading to improved accuracy and robustness.

## Contribution

Introduces hidden two-stream networks for motion representation without optical flow and a video prediction-based data augmentation framework for semantic segmentation.

## Key findings

- Enhanced action recognition accuracy without optical flow
- More robust semantic segmentation through data augmentation
- Frameworks outperform previous models in experiments

## Abstract

In this dissertation, I present my work towards exploring temporal information for better video understanding. Specifically, I have worked on two problems: action recognition and semantic segmentation. For action recognition, I have proposed a framework, termed hidden two-stream networks, to learn an optimal motion representation that does not require the computation of optical flow. My framework alleviates several challenges faced in video classification, such as learning motion representations, real-time inference, multi-framerate handling, generalizability to unseen actions, etc. For semantic segmentation, I have introduced a general framework that uses video prediction models to synthesize new training samples. By scaling up the training dataset, my trained models are more accurate and robust than previous models even without modifications to the network architectures or objective functions. I believe videos have much more potential to be mined, and temporal information is one of the most important cues for machines to perceive the visual world better.