TFPose: Direct Human Pose Estimation with Transformers
This addresses pose estimation for computer vision applications, offering a competitive regression-based alternative to heatmap methods, though it is incremental in approach.
The paper tackles human pose estimation by reformulating it as a sequence prediction problem using transformers, bypassing heatmap-based drawbacks and improving feature alignment, achieving results comparable to state-of-the-art heatmap-based methods on MS-COCO and MPII datasets.
We propose a human pose estimation framework that solves the task in the regression-based fashion. Unlike previous regression-based methods, which often fall behind those state-of-the-art methods, we formulate the pose estimation task into a sequence prediction problem that can effectively be solved by transformers. Our framework is simple and direct, bypassing the drawbacks of the heatmap-based pose estimation. Moreover, with the attention mechanism in transformers, our proposed framework is able to adaptively attend to the features most relevant to the target keypoints, which largely overcomes the feature misalignment issue of previous regression-based methods and considerably improves the performance. Importantly, our framework can inherently take advantages of the structured relationship between keypoints. Experiments on the MS-COCO and MPII datasets demonstrate that our method can significantly improve the state-of-the-art of regression-based pose estimation and perform comparably with the best heatmap-based pose estimation methods.