[Retracted] A Study of Athlete Pose Estimation Techniques in Sports Game Videos Combining Multiresidual Module Convolutional Neural Networks

Abstract

In this paper, we propose a multiresidual module convolutional neural network-based method for athlete pose estimation in sports game videos. The network firstly designs an improved residual module based on the traditional residual module. Firstly, a large perceptual field residual module is designed to learn the correlation between the athlete components in the sports game video within a large perceptual field. A multiscale residual module is designed in the paper to better solve the inaccuracy of the pose estimation due to the problem of scale change of the athlete components in the sports game video. Secondly, these three residual modules are used as the building blocks of the convolutional neural network. When the resolution is high, the large perceptual field residual module and the multiscale residual module are used to capture information in a larger range as well as at each scale, and when the resolution is low, only the improved residual module is used. Finally, four multiresidual module convolutional neural networks are used to form the final multiresidual module stacked convolutional neural network. The neural network model proposed in this paper achieves high accuracy of 89.5% and 88.2% on the upper arm and lower arm, respectively, so the method in this paper reduces the influence of occlusion on the athlete’s posture estimation to a certain extent. Through the experiments, it can be seen that the proposed multiresidual module stacked convolutional neural network-based method for athlete pose estimation in sports game videos further improves the accuracy of athlete pose estimation in sports game videos.

Research on Multiplayer Posture Estimation Technology of Sports Competition Video Based on Graph Neural Network Algorithm

“…In particular, for the wrist joint points most concerned in the process of human posture estimation, the positioning accuracy of this algorithm is significantly higher than that of the other four methods. For example, taking 30 pixels as the threshold, the wrist PDJ value of this algorithm is 98.3%, which far exceeds the PDJ value of reference [20] 82.1%, reference [21] 84.1%, reference [22] 78.2% and reference [23] 80.1%. Ours algorithm [20] algorithm [21] algorithm [22] algorithm [23] Figure 8: PDJ curve comparing this algorithm with other algorithms on VIPS-VideoPose database.…”

“…For example, taking 30 pixels as the threshold, the wrist PDJ value of this algorithm is 98.3%, which far exceeds the PDJ value of reference [20] 82.1%, reference [21] 84.1%, reference [22] 78.2% and reference [23] 80.1%. Ours algorithm [20] algorithm [21] algorithm [22] algorithm [23] Figure 8: PDJ curve comparing this algorithm with other algorithms on VIPS-VideoPose database.…”

“…e human posture estimation process is divided into two stages: detection component and positioning component [20]. In the first stage, the specific deep learning network is used as the human part detector to realize the detection function of human parts; In the second stage, a new spatial constraint relationship of human parts is designed to uniquely determine the position of each part from the candidate box of human parts, realize the positioning function of joint points, and finally complete the estimation process of human posture.…”

“…In order to test the running time of this algorithm in the process of 3D human body pose estimation, this algorithm is compared with literature [20] and literature [21] under the same software and hardware conditions. e results are shown in Figure 7.…”

“…e results are shown in Figure 7. e average time per frame of algorithm [20] is 32.801 s, because the model mainly matches the SMPL model with two-dimensional joint points based on parameter optimization, and the optimization process takes a lot of time; SMPL needs to consider not only the basic human body configuration, but also whether the joint density can control the movement of points. erefore, when there are no obvious joints in the chest, several joints are added, while the number of joints on the hand is less.…”