We propose a new method based on local space-time interest points and self-organization feature maps(SOFM)to recognize and retrieval complex events in real movie.In this method

an individual video sequence is represented as a SOFM density map

We integrate this density map with a support vector machine(SVM)to recognize events.Local space-time features are introduced to capture the local events in video and can be adapted to size and velocity of the pattern of the event.To evaluate the effectiveness of this method

we use the public Hollywood dataset.In this dataset shot sequences are collected from 32 different Hollywood movies and it includes eight event classes.According to the experiment

the average accuracy rate

the average precision rate

and average recall rate were 0.601

0.530 and 0.566 respectively.The presented results justify the proposed method explicitly improving the average accuracy and average precision compared with other relative approaches.