Pixel-Coordinate-Induced Human Pose High-Precision Estimation Method

Accurately estimating human pose is crucial for providing feedback during exercises or musical performances, but the complex and flexible nature of joints makes it challenging. Additionally, traditional methods often neglect pixel coordinates, which are naturally present in high-resolution images body. To address this issue, we propose a novel estimation method that directly incorporates coordinates. Our adds coordinate channel to convolution process embeds coordinates into feature map, while also using attention capture position- structure-sensitive features. We further reduce network parameters computational cost by small-scale kernels smooth activation function residual blocks. evaluate our model on MPII Human Pose COCO Keypoint Detection datasets demonstrate improved accuracy, highlighting importance incorporating location information position-sensitive tasks.