Actions and Attributes from Wholes and Parts
This work addresses computer vision tasks like action and attribute classification, offering incremental improvements by integrating part-based methods with existing holistic approaches.
The paper tackles action and attribute classification by developing a part-based approach using deep poselet detectors and convolutional neural networks, achieving top-performing results on both tasks and demonstrating effectiveness with state-of-the-art person detection systems instead of oracle detectors.
We investigate the importance of parts for the tasks of action and attribute classification. We develop a part-based approach by leveraging convolutional network features inspired by recent advances in computer vision. Our part detectors are a deep version of poselets and capture parts of the human body under a distinct set of poses. For the tasks of action and attribute classification, we train holistic convolutional neural networks and show that adding parts leads to top-performing results for both tasks. In addition, we demonstrate the effectiveness of our approach when we replace an oracle person detector, as is the default in the current evaluation protocol for both tasks, with a state-of-the-art person detection system.