Adversarial Semantic Data Augmentation for Human Pose Estimation

Abstract

Human pose estimation is the task of localizing body keypoints from still images. The state-of-the-art methods suffer from insufficient examples of challenging cases such as symmetric appearance, heavy occlusion and nearby person. To enlarge the amounts of challenging cases, previous methods augmented images by cropping and pasting image patches with weak semantics, which leads to unrealistic appearance and limited diversity. We instead propose Semantic Data Augmentation (SDA), a method that augments images by pasting segmented body parts with various semantic granularity. Furthermore, we propose Adversarial Semantic Data Augmentation (ASDA), which exploits a generative network to dynamiclly predict tailored pasting configuration. Given offthe-shelf pose estimation network as discriminator, the generator seeks the most confusing transformation to increase the loss of the discriminator while the discriminator takes the generated sample as input and learns from it. The whole pipeline is optimized in an adversarial manner. State-of-the-art results are achieved on challenging benchmarks. The code has been publicly available at https://github.com/Binyr/ASDA.

EvoPose2D: Pushing the Boundaries of 2D Human Pose Estimation Using Accelerated Neuroevolution With Weight Transfer

“…An issue surrounding the 2D human pose estimation literature is that it is often difficult to make fair comparisons of model performance due to the heavy use of model-agnostic improvements. Examples include the use of different learning rate schedules [24], [49], more data augmentation [49], [50], loss functions that target more challenging keypoints [23], specialized post-processing steps [51], [52], or more accurate person detectors [49], [52]. These discrepancies in training algorithms can potentially account for the reported differences in accuracy.…”

“…. , 0.90, 0.95), AP 50 (AP at OKS = 0.50), AP 75 , AP M (medium objects), AP L (large objects), and AR (mean AR at OKS = 0.50, 0.55, . .…”

Self-Supervised Learning for Human Pose Estimation in Sports

“…Human pose estimation is a field of active interest in computer vision. The currently best scoring approaches on common benchmarks like COCO [4] or MPII Human Pose [5] are based on convolutional neural networks [6,7]. In contrast to multi-stage approaches like Mask R-CNN [8] which find person bounding boxes at first and detect one keypoint of each type in the second step, these recent approaches find all keypoints in a single step.…”

“…In contrast to multi-stage approaches like Mask R-CNN [8] which find person bounding boxes at first and detect one keypoint of each type in the second step, these recent approaches find all keypoints in a single step. The underlying backbone used in many recent models like [6,7] is the High Resolution Net (HRNet) [9] and its variants for human pose estimation, e.g. [10].…”

Multi-Cascade Perceptual Human Posture Recognition Enhancement Network

“…In addition to improving the resolution of the hot spot map in the classification sub-network in different ways, Yanrui Bin et al [58] proposed Semantic Data Augmentation (SDA). By pasting segmented body parts with various semantic granularity, it can enhance the effect of human body detection in the image when the target in the detected image is seriously occluded.…”