关键词: 直方图修正, 条件概率, 图像增强, RBF 神经网络

Abstract: For low-contrast image enhancement problem, we propose an algorithm based on histogram
correction and RBF neural network methods. Obtained the conditional probability histogram of the
pixels in the presence of contrast with its neighborhood through original image, adjusting the weights
of two parameters can change the conditional probability histogram and uniform distribution
histogram. In this paper, RBF neural network is applied to set up the nonlinear mapping between
image features and two enhanced parameters. In order to achieve adaptive image enhancement, rapid
enhancement parameters are obtained according to the characteristics of the original image. The
results show this method has good real-time ability, wide range of application, low computational
complexity and good adaptability.

Key words: histogram modification, conditional probability, image enhancement, RBF neural network