An Adaptive Image Enhancement Based on the Vector Closed Operations

This paper presents an adaptive image enhancement algorithm based on the vector closed operations, aiming at distinguishing features from impulse noise and salt & pepper noise, and solving the problem of the inconvenience in interactively modifying the parameter to adjust the contrast. According to the characteristic that grey level values arising from noise are normally weakly correlated and those arising from features are strongly correlated, the proposed algorithm can distinguish features from noise in each scale using correlation values, and then they can be adoptively processed independently. Therefore, the algorithm can adoptively enhance more features and suppress more noise. Meanwhile it can adoptively adjust the contrast based on the visual statistics theory. After the contrast is adjusted, the mean value of the enhanced image is around the gray of the visually optimized image. Therefore, the finally enhanced image has good sharpening features with noise suppressed and contrast moderately enhanced.