S Rajaprabha, M Sugadev
2016 unpublished
Many promising applications need to track the blurred images in the videos. In most of the object trackers, implicit assumptions are made that the video is blur free. However, in real time videos, motion blur is very common. If severe blur is present in a video the performance of the generic object trackers may go down significantly. The proposed method uses GLCM algorithm for feature extractions from the blurred object and then ANFIS (Adaptive Neuro Fuzzy Inference System) for tracking those
more » ... or tracking those blurred objects in the videos. The ANFIS model is trained with the parameters of blurred objects. The input video is imported and GLCM (Gray-Level Co occurrence Matrix) method is used to extract features from the frame. Now the ANFIS data is loaded and compared with the frame. Then the blurred object is detected and tracked by the ANFIS model. The proposed algorithm robustly tracks challenging scenes and severely blurred videos. The speed and performance is improved in this proposed method.