Temporal Pyramid Network for Action Recognition

Abstract

Visual tempo characterizes the dynamics and the temporal scale of an action. Modeling such visual tempos of different actions facilitates their recognition. Previous works often capture the visual tempo through sampling raw videos at multiple rates and constructing an input-level frame pyramid, which usually requires a costly multi-branch network to handle. In this work we propose a generic Temporal Pyramid Network (TPN) at the feature-level, which can be flexibly integrated into 2D or 3D backbone networks in a plug-and-play manner. Two essential components of TPN, the source of features and the fusion of features, form a feature hierarchy for the backbone so that it can capture action instances at various tempos. TPN also shows consistent improvements over other challenging baselines on several action recognition datasets. Specifically, when equipped with TPN, the 3D ResNet-50 with dense sampling obtains a 2% gain on the validation set of Kinetics-400. A further analysis also reveals that TPN gains most of its improvements on action classes that have large variances in their visual tempos, validating the effectiveness of TPN.

arXiv commentTo appear in CVPR 2020. Code is available at https://github.com/decisionforce/TPN

References in corpus (14)

• Very Deep Convolutional Networks for Large-Scale Image Recognition
• Batch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Shift
• Improving neural networks by preventing co-adaptation of feature detectors
• Two-Stream Convolutional Networks for Action Recognition in Videos
• Accurate, Large Minibatch SGD: Training ImageNet in 1 Hour
• Action Recognition with Trajectory-Pooled Deep-Convolutional Descriptors
• Feature Pyramid Networks for Object Detection
• Convolutional Two-Stream Network Fusion for Video Action Recognition
• Temporal Segment Networks: Towards Good Practices for Deep Action Recognition
• Scaling Egocentric Vision: The EPIC-KITCHENS Dataset
• Video Classification with Channel-Separated Convolutional Networks
• Temporal Relational Reasoning in Videos
• Intra- and Inter-Action Understanding via Temporal Action Parsing
• Dynamic Temporal Pyramid Network: A Closer Look at Multi-Scale Modeling for Activity Detection

Cited by in corpus (5)

• Human Action Recognition from Various Data Modalities: A Review
• GSRFormer: Grounded Situation Recognition Transformer with Alternate Semantic Attention Refinement
• RoVISQ: Reduction of Video Service Quality via Adversarial Attacks on Deep Learning-based Video Compression
• Global Information Guided Video Anomaly Detection
• Multi-Label Activity Recognition using Activity-specific Features and Activity Correlations