Research On Saliency Object Detection Method For Object Region Fusion

The saliency object detection method processes the image to be detected,and the generated saliency map has inaccurate detection of the target area,that is,redundant pixel information is used as the domain information of the target area,the boundary definition of the target area is insufficient,and the details in the target area are not obvious.That is,the lack of contrast or the inadequacy of the related algorithm itself due to filtering reasons.Aiming at the above problems,a significant target detection method for target region fusion is proposed.First,the color and brightness features of the extracted image in the Lab color space are used as a basis for the significance value to generate a saliency map.The power law transformation function is used to reduce the contrast between the target area and the background area in the saliency map to suppress the background of the saliency map from highlighting the target area.Secondly,the saliency map with good background suppression effect is binarized to simplify the calculation of the program.The generated binary image is used as a segmentation template to multiply the original image,and only highlights the original image target region.According to the comparison of the average gradient of the saliency map generated by the IG and LC algorithms,the LC significant target detection method is used to generate a clear target region.Figure.At the same time,based on the transparency value in the map equation and the contour of the artificial background region,the transparency of the pixel in the background region is set to zero to suppress the background region of the LC algorithm to highlight the target region of the significant image.The target area of the salient map is enhanced by the image enhancement of the adaptive fireworks algorithm to enhance the contrast of the saliency map.Finally,the two target regions with good background suppression and clear detail are weighted and linearly fused to generate the final saliency map.Using the standard data set to test the detection method and compare it with the more popular saliency target detection method,the results show that the saliency map generated by the method has the characteristics of clear target area and clear details of the target area.And has better robustness.The paper has 42 figures,3 tables and 53 references.