TasksSotADatasetsPapersMethodsSubmitAbout
Papers With Code 2

A community resource for machine learning research: papers, code, benchmarks, and state-of-the-art results.

Explore

Notable BenchmarksAll SotADatasetsPapersMethods

Community

Submit ResultsAbout

Data sourced from the PWC Archive (CC-BY-SA 4.0). Built by the community, for the community.

Papers/Rethinking the Faster R-CNN Architecture for Temporal Acti...

Rethinking the Faster R-CNN Architecture for Temporal Action Localization

Yu-Wei Chao, Sudheendra Vijayanarasimhan, Bryan Seybold, David A. Ross, Jia Deng, Rahul Sukthankar

2018-04-20CVPR 2018 6Action ClassificationAction LocalizationGeneral ClassificationTemporal Action Localizationobject-detectionObject Detection
PaperPDF

Abstract

We propose TAL-Net, an improved approach to temporal action localization in video that is inspired by the Faster R-CNN object detection framework. TAL-Net addresses three key shortcomings of existing approaches: (1) we improve receptive field alignment using a multi-scale architecture that can accommodate extreme variation in action durations; (2) we better exploit the temporal context of actions for both proposal generation and action classification by appropriately extending receptive fields; and (3) we explicitly consider multi-stream feature fusion and demonstrate that fusing motion late is important. We achieve state-of-the-art performance for both action proposal and localization on THUMOS'14 detection benchmark and competitive performance on ActivityNet challenge.

Results

TaskDatasetMetricValueModel
VideoTHUMOS’14Avg mAP (0.3:0.7)39.8TAL-Net
VideoTHUMOS’14mAP IOU@0.159.8TAL-Net
VideoTHUMOS’14mAP IOU@0.257.1TAL-Net
VideoTHUMOS’14mAP IOU@0.353.2TAL-Net
VideoTHUMOS’14mAP IOU@0.448.5TAL-Net
VideoTHUMOS’14mAP IOU@0.542.8TAL-Net
VideoTHUMOS’14mAP IOU@0.633.8TAL-Net
VideoTHUMOS’14mAP IOU@0.720.8TAL-Net
Temporal Action LocalizationTHUMOS’14Avg mAP (0.3:0.7)39.8TAL-Net
Temporal Action LocalizationTHUMOS’14mAP IOU@0.159.8TAL-Net
Temporal Action LocalizationTHUMOS’14mAP IOU@0.257.1TAL-Net
Temporal Action LocalizationTHUMOS’14mAP IOU@0.353.2TAL-Net
Temporal Action LocalizationTHUMOS’14mAP IOU@0.448.5TAL-Net
Temporal Action LocalizationTHUMOS’14mAP IOU@0.542.8TAL-Net
Temporal Action LocalizationTHUMOS’14mAP IOU@0.633.8TAL-Net
Temporal Action LocalizationTHUMOS’14mAP IOU@0.720.8TAL-Net
Zero-Shot LearningTHUMOS’14Avg mAP (0.3:0.7)39.8TAL-Net
Zero-Shot LearningTHUMOS’14mAP IOU@0.159.8TAL-Net
Zero-Shot LearningTHUMOS’14mAP IOU@0.257.1TAL-Net
Zero-Shot LearningTHUMOS’14mAP IOU@0.353.2TAL-Net
Zero-Shot LearningTHUMOS’14mAP IOU@0.448.5TAL-Net
Zero-Shot LearningTHUMOS’14mAP IOU@0.542.8TAL-Net
Zero-Shot LearningTHUMOS’14mAP IOU@0.633.8TAL-Net
Zero-Shot LearningTHUMOS’14mAP IOU@0.720.8TAL-Net
Action LocalizationTHUMOS’14Avg mAP (0.3:0.7)39.8TAL-Net
Action LocalizationTHUMOS’14mAP IOU@0.159.8TAL-Net
Action LocalizationTHUMOS’14mAP IOU@0.257.1TAL-Net
Action LocalizationTHUMOS’14mAP IOU@0.353.2TAL-Net
Action LocalizationTHUMOS’14mAP IOU@0.448.5TAL-Net
Action LocalizationTHUMOS’14mAP IOU@0.542.8TAL-Net
Action LocalizationTHUMOS’14mAP IOU@0.633.8TAL-Net
Action LocalizationTHUMOS’14mAP IOU@0.720.8TAL-Net

Related Papers

A Real-Time System for Egocentric Hand-Object Interaction Detection in Industrial Domains2025-07-17RS-TinyNet: Stage-wise Feature Fusion Network for Detecting Tiny Objects in Remote Sensing Images2025-07-17Decoupled PROB: Decoupled Query Initialization Tasks and Objectness-Class Learning for Open World Object Detection2025-07-17Dual LiDAR-Based Traffic Movement Count Estimation at a Signalized Intersection: Deployment, Data Collection, and Preliminary Analysis2025-07-17DVFL-Net: A Lightweight Distilled Video Focal Modulation Network for Spatio-Temporal Action Recognition2025-07-16Vision-based Perception for Autonomous Vehicles in Obstacle Avoidance Scenarios2025-07-16Tomato Multi-Angle Multi-Pose Dataset for Fine-Grained Phenotyping2025-07-15ECORE: Energy-Conscious Optimized Routing for Deep Learning Models at the Edge2025-07-08