The improved two-dimensional Gabor filter based interest objects detection

Saliency using natural statistics model has a good effect on obtaining the region of interest. However, it still has limitations such as insufficient image contrast, inability to cope with local information and so on. In order to overcome these challenges, an improved algorithm based on the two-dimensional Gabor filter and color maps is proposed. The two-dimensional Gabor filter has a close approach to simple receptive fields in human striate cortex. It is used to tackle the problem mentioned above. Experiments have demonstrated the feasibility and effectiveness of the proposed algorithm for interest objects detection.