Image enhancement using multi scale image features extracted by top-hat transform

To efficiently enhance images, a novel algorithm using multi scale image features extracted by top-hat transform is proposed in this paper. Firstly, the multi scale bright and dim regions are extracted through top-hat transform using structuring elements with the same shape and increasing sizes. Then, two types of multi scale image features, which are the multi scale bright and dim image regions at each scale and the multi scale image details between neighboring scales, are extracted and used to form the final extracted bright and dim image regions. Finally, the image is enhanced through enlarging the contrast between the final extracted bright and dim image features. Experimental results on images from different applications verified that the proposed algorithm could efficiently enhance the contrast and details of image, and produce few noise regions.

► This paper proposes a novel multi scale top-hat transform based algorithm.
► The algorithm extracts and uses two types of multi scale image features.
► The first is multi scale image regions at each scale.
► The second is multi scale image details between neighboring scales.
► The algorithm enhances both the contrast and image details, and produces few noise regions.