Late Temporal Modeling in 3D CNN Architectures with BERT for Action Recognition

M. Esat Kalfaoglu, Sinan Kalkan, A. Aydin Alatan

2020-08-03Action Recognition

Abstract

In this work, we combine 3D convolution with late temporal modeling for action recognition. For this aim, we replace the conventional Temporal Global Average Pooling (TGAP) layer at the end of 3D convolutional architecture with the Bidirectional Encoder Representations from Transformers (BERT) layer in order to better utilize the temporal information with BERT's attention mechanism. We show that this replacement improves the performances of many popular 3D convolution architectures for action recognition, including ResNeXt, I3D, SlowFast and R(2+1)D. Moreover, we provide the-state-of-the-art results on both HMDB51 and UCF101 datasets with 85.10% and 98.69% top-1 accuracy, respectively. The code is publicly available.

Results

Task	Dataset	Metric	Value	Model
Activity Recognition	HMDB-51	Average accuracy of 3 splits	85.1	R2+1D-BERT
Activity Recognition	UCF 101	3-fold Accuracy	98.69	R2+1D-BERT
Action Recognition	HMDB-51	Average accuracy of 3 splits	85.1	R2+1D-BERT
Action Recognition	UCF 101	3-fold Accuracy	98.69	R2+1D-BERT

Related Papers

A Real-Time System for Egocentric Hand-Object Interaction Detection in Industrial Domains2025-07-17 Zero-shot Skeleton-based Action Recognition with Prototype-guided Feature Alignment2025-07-01 EgoAdapt: Adaptive Multisensory Distillation and Policy Learning for Efficient Egocentric Perception2025-06-26 Feature Hallucination for Self-supervised Action Recognition2025-06-25 CARMA: Context-Aware Situational Grounding of Human-Robot Group Interactions by Combining Vision-Language Models with Object and Action Recognition2025-06-25 Including Semantic Information via Word Embeddings for Skeleton-based Action Recognition2025-06-23 Adapting Vision-Language Models for Evaluating World Models2025-06-22 Active Multimodal Distillation for Few-shot Action Recognition2025-06-16