Selection Of The Suitable Receiving Parts By Deep Learning-Based Feature Image Analysis Method

To comprehensively understand the variations in joints under different suitable ball-receiving situations of football. A deep learning-based K-means clustering feature image analysis algorithm is constructed to study the variations of joints in three different receiving parts of football: (1) while receiving the ball in the air with the medial side of foot, the displacements and angles of knee, hip, and ankle joints; (2) while receiving the bounce ball with the medial side of foot, the displacements and angles of knee, hip, and ankle joints; (3) while receiving the ball in the air with the backside of foot, the displacements and angles of knee, hip, and ankle joints. By qualifying the displacements and angles of the above joints, the accuracy of selecting the receiving parts can be determined. The results show that the deep learning K-means clustering feature image analysis algorithm is accurate in selecting the knee, hip, and ankle joint displacements and angles under different ball-oncoming situations. As for the displacement values of joints while re ceiving the bouncing ball with the medial side of the foot, receiving the ball in the air with the medial side of the foot, and receiving the ball in the air with the backside of foot, the judgments are close to the standard values. Also, the algorithm shows excellent stability while selecting and calculating the receiving parts. Based on the deep learning K-means clustering feature image analysis, the displacements and angles in the knee, hip, and ankle joints under three different ball-receiving situations are judged and selected, which provides suggestions for football training and promotion.