At present days most the researchers are carried out a lot of color enhancement of gray images in diverse fields. It is commonly used methods mainly include density layering method gray scale color transformation method pixel self-transformation method rainbow coding, method, metal coding as well as pseudo color enhancement algorithm based on frequency domain. In current centuries signal as well as image processing is based on fractional calculus has attracted extensive attention. Color Enhancement is realized by utilizing the constructed high gray scale enhancement algorithm. Combined with the convolution neural network, extract the features of the multiscale image utilizing the compact learning method. The research acquisition show that the compared with the traditional image enhancement image proposed method give better comprehensive performance in subjective vision besides objective indicators in dealing with low light image enhancement. The brightness distribution of the enhanced image can well restore the brightness distribution of the real shooting environment in addition to higher robustness. In this research paper the Retinex Hetergenous equalization Feature Fusion Scale Image Enhancement method for content based deep learning enhancement system. The proposed method effectively with traditional jet coding as well as HSV pseudo color methods as well as find out the brightness of image, brightness of the distortion, blocking can well restore the brightness besides preserves the values of the RGB color components of each pixel of the image in addition to estimates the reflectance distribution of the real shooting environment has higher robustness