A Multi-Scale Image Enhancement Model using Human Visual System Characteristics

Image enhancement is a fundamental preprocessing step for many automated systems and vision systems. Many enhancement algorithms have been anticipated based on different sets of criteria. One of the most widely used algorithms is the direct multi-scale image enhancement algorithm. The specialty of this algorithm is, it provides contrast enhancement, tonal rendition, dynamic range compression and accurate edge preservation of the images. It also provides these features to the individual images and/or simultaneously to the images. In this proposed method, a multi-scale image enhancement algorithm is established by using parametric contrast measure with the transform techniques such as Laplacian pyramid, discrete wavelet transform, Stationary wavelet transform and Dual-tree complex wavelet transform. The new contrast measure provides both the luminance and contrast masking characteristics of the human visual system. The proposed method is used to attain simultaneous local and global enhancements. The enhancement measures such as Entropy, Mean opinion score and Measure of enhancement gives better results than the existing methods.