Abstract

Background and Objective: medicine is one of the applied areas of image processing that is currently widely applied in the analysis of images obtained from medical imaging devices, CT-scan and MRI included. Slight differences in an image’s contrast can bring about considerable effects on the image quality. The present study attends to the improvement of medical images’ contrast as one of the necessary preprocessing stages. Study Method: the algorithm proposed in the present study tries to reduce the effect of uneven light so that the maximum details of an image can be preserved. The algorithm proposed in the current research paper has been simulated using MATLAB. The dataset pertaining to brain CT-scan and MRI has been utilized herein. A sample of a brain MRI image is the input to the proposed algorithm. The image has been analyzed by the use of Poisson relations and equations and the final improved image will be exhibited. Then, the proposed algorithm will be compared with neural network and genetic algorithm. Findings: according to the results of comparing the efficiency of the proposed algorithm with those of the genetic algorithm and back-propagation neural network, it can be clearly discerned that, besides the notable speed of the proposed algorithm, it also offers defendable results versus back-propagation neural network but it is found not much successful in contrast to genetic-based algorithm. Conclusion: screened Poisson equation is an effective and, in the meantime, fast method for improving the contrast of the medical images, especially tomography images of the soft tissues like the cerebral tissue.