The Bowl-Throwing Result Analysis for Bowling Game

In this paper, we propose a new approach to bowling game result analysis. A reliable attribute, i.e. the lane boundary, is detected first. Based on this attribute the rudimentary result frame can be detected. The spectral features of the audio frames of the candidates are employed to determine the exact result frame by applying the DistAl algorithm, a simple and fast constructive neural-network learning algorithm for pattern classification. The color information of the pin mark is then utilized to detect the preliminary bowling pin position. According to the pin arrangement rule and the proposed pin-mark curve the exact position for each pin can be decided. Hence, the bowling game result for each bowl can be indexed. Our experimental results show that the proposed approach is effective and robust.