Non-uniform illumination correction in infrared images based on a modified fuzzy c-means algorithm

The correction of non-uniform illumination and the elimination of shading artifacts is an important preprocessing task used in a great number of image processing applications such as segmentation, registration or quantitative analysis. Although, a careful and accurate set up of the image acquisition system may degrade the importance of a brightness normalization algorithm, non-uniform illumination appears due to the interaction of objects and light on the scene requires retrospective shading correction. The image formation process and the corresponding shading effects are described by a linear image formation model, which consists of a multiplicative and an additive shading component. In this paper a novel brightness normalization method is proposed to eliminate the non-uniform illumination effects. The method is based on the application of a fuzzy c-means algorithm (FCM) only on the background part of the acquired image, where the objective function is modified to take into account local information of each pixel in the estimation of the multiplicative and the additive shading components. The modified FCM algorithm is iterative, as the standard FCM, and at each iteration the multiplicative and the additive shading components are re-estimated based on the cluster centers and the membership of each pixel in a specific cluster. Brightness correction is performed by the inverse of the image formation model after FCM convergence. Experiments were conducted in a database of both real and artificial infrared images. The experimental results show that the proposed method decreases significantly the non-uniform illumination effects and does not introduce brightness variations if the background is uniform.