Learning Deep Feature Representations with Domain Guided Dropout for Person Re-identification
This work addresses the domain-specific problem of person re-identification by enabling better feature learning across multiple datasets, though it appears incremental as it builds on existing CNN methods with a novel dropout technique.
The paper tackles the problem of learning robust feature representations from multiple domains for person re-identification, where individual datasets are insufficient in size, and proposes a Domain Guided Dropout algorithm that improves feature learning, resulting in methods that outperform state-of-the-art approaches on multiple datasets by large margins.
Learning generic and robust feature representations with data from multiple domains for the same problem is of great value, especially for the problems that have multiple datasets but none of them are large enough to provide abundant data variations. In this work, we present a pipeline for learning deep feature representations from multiple domains with Convolutional Neural Networks (CNNs). When training a CNN with data from all the domains, some neurons learn representations shared across several domains, while some others are effective only for a specific one. Based on this important observation, we propose a Domain Guided Dropout algorithm to improve the feature learning procedure. Experiments show the effectiveness of our pipeline and the proposed algorithm. Our methods on the person re-identification problem outperform state-of-the-art methods on multiple datasets by large margins.