The supporting system for assessment of catwalk modeling using variable module graph method based on video
暂无分享,去创建一个
The previous research with the title Analysis and Simulation of Gait Quality Assessment System for Modeling School Based on Video Processing Using Variable Module Graph Method has obtained a system accuracy rates of 52%. The current research is conducted by adding extra features and making it into several testing scenarios to see which of those give the highest accuracy rates compared to the previous research. Those features addition are the segmentation of the Zero Crossing Graph and the combination with the average changes of object velocity from each sample. With those features addition make the performance of a system increased into 56.76% from the scenario of combining the Zero Crossing 4-segmentation with average changes of object velocity features.
[1] Thomas S. Huang,et al. Event Detection Using "Variable Module Graphs" for Home Care Applications , 2007, EURASIP J. Adv. Signal Process..
[2] Thomas S. Huang,et al. Variable module graphs: a framework for inference and learning in modular vision systems , 2005, IEEE International Conference on Image Processing 2005.