Soccer Video Shot Classification Method Based on Color and Edge Distribution

Classified by the characters of shots for soccer video,a color and edge distribution based shot classification method is presented.A Gaussian mixture model(GMM) is first trained for grass pixel,the grass distribution is then computed based on the trained GMM.Due to the sensitivity to lights for color features,edge distribution is computed to complement the drawbacks of color features.Integrating the two types of features embodies the main characters of different shot types.Support vector machines(SVMs) which suit to classification problem for tiny samples is designed for different shot types through the features extracted by the method.Experiments showed that the average precision of the method can attain over 92%.