Unsupervised Learning of Video Representations using LSTMs

TL;DR

This paper introduces an unsupervised approach using LSTM networks to learn video representations, which can be used for various tasks including sequence reconstruction, future prediction, and improving action recognition, especially with limited labeled data.

Contribution

The paper presents a novel unsupervised learning framework with LSTMs for video representation that enhances downstream tasks like action recognition, even with out-of-domain pretraining.

Findings

LSTM-based representations improve action recognition accuracy.

Pretraining on large unlabeled video data benefits supervised tasks.

The model can extrapolate and visualize learned video features.

Abstract

We use multilayer Long Short Term Memory (LSTM) networks to learn representations of video sequences. Our model uses an encoder LSTM to map an input sequence into a fixed length representation. This representation is decoded using single or multiple decoder LSTMs to perform different tasks, such as reconstructing the input sequence, or predicting the future sequence. We experiment with two kinds of input sequences - patches of image pixels and high-level representations ("percepts") of video frames extracted using a pretrained convolutional net. We explore different design choices such as whether the decoder LSTMs should condition on the generated output. We analyze the outputs of the model qualitatively to see how well the model can extrapolate the learned video representation into the future and into the past. We try to visualize and interpret the learned features. We stress test the model by running it on longer time scales and on out-of-domain data. We further evaluate the representations by finetuning them for a supervised learning problem - human action recognition on the UCF-101 and HMDB-51 datasets. We show that the representations help improve classification accuracy, especially when there are only a few training examples. Even models pretrained on unrelated datasets (300 hours of YouTube videos) can help action recognition performance.