Remarks on human body posture estimation using neural network and Kalman filter

This paper proposes a human body posture estimation method from image using ANN. The input feature vector of the ANN is composed with the result of analyzing a human silhouette image and the output vector of the ANN indicates the 2D coordinates of the human body´s significant points, such as head, shoulders, hands, elbows, knees, and feet. The estimated results are optimized and tracked by using Kalman filter. The proposed estimation method is implemented on a personal computer and runs in real-time. Experimental results show both the feasibility and the effectiveness of the proposed method for estimating human body postures.