R-CNNs for Pose Estimation and Action Detection

TL;DR

This paper introduces R-CNN based methods for pose estimation and action detection in images, achieving state-of-the-art results and providing a new dataset for action localization and classification.

Contribution

The paper develops R-CNN models tailored for keypoint and action prediction, and introduces a new dataset for joint localization and action classification.

Findings

State-of-the-art results on PASCAL VOC for pose and action prediction

Effective R-CNN training with task-specific loss functions

New dataset for action detection with promising baseline results

Abstract

We present convolutional neural networks for the tasks of keypoint (pose) prediction and action classification of people in unconstrained images. Our approach involves training an R-CNN detector with loss functions depending on the task being tackled. We evaluate our method on the challenging PASCAL VOC dataset and compare it to previous leading approaches. Our method gives state-of-the-art results for keypoint and action prediction. Additionally, we introduce a new dataset for action detection, the task of simultaneously localizing people and classifying their actions, and present results using our approach.