Actions and Attributes from Wholes and Parts
arXiv:1412.2604
Abstract
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.
References in corpus (7)
- Very Deep Convolutional Networks for Large-Scale Image Recognition
- Caffe: Convolutional Architecture for Fast Feature Embedding
- DeepPose: Human Pose Estimation via Deep Neural Networks
- Rich feature hierarchies for accurate object detection and semantic segmentation
- Bird Species Categorization Using Pose Normalized Deep Convolutional Nets
- Deformable Part Models are Convolutional Neural Networks
- Deep Poselets for Human Detection
Cited by in corpus (9)
- Very Deep Convolutional Networks for Large-Scale Image Recognition
- Contextual Action Recognition with R*CNN
- Pedestrian Attribute Recognition: A Survey
- Weakly and Semi Supervised Human Body Part Parsing via Pose-Guided Knowledge Transfer
- Part-Stacked CNN for Fine-Grained Visual Categorization
- DeepVoting: A Robust and Explainable Deep Network for Semantic Part Detection under Partial Occlusion
- Deep Imbalanced Attribute Classification using Visual Attention Aggregation
- Objects as context for detecting their semantic parts
- Learning Actor Relation Graphs for Group Activity Recognition