Adaptive Color Correction and Contrast Enhancement Algorithm for Low Illumination Images

Illumination changes will cause color distortion and contrast reduction of images.In order to make computer vision system have constant color sensing function,this paper presents a color correction and contrast enhancement algorithm based on extreme learning machine(ELM)and cuckoo search algorithm.Firstly,using a 17 dimensional feature vector extracted from an input image as the input of a trained ELM,it selected the optimal type of the optimal color constancy algorithm based on the trained ELM adaptively,and then it corrected the color of the image by the selected color constancy algorithm.Secondly,it used the Cuckoo search algorithm to select the optimal parameters of brightness enhancement function for the light intensity component of the image,and then it enhanced the light intensity image by the corresponding enhancement function.The experimental results based on a large benchmark dataset named Funt show that the proposed algorithm can not only correct the color of an image effectively,but also enhance the information and contrast of the image adaptively,thereby obtaining the optimal visual pleasing effect of the image both in color and in resolution.