A framework for detection of jaundice in new born babies using homomorphic filtering based image processing

摘要

Digital Image processing in medical field can play an important role in monitoring and diagnosis the patient's health condition to provide an effective treatment. However, quality of camera, luminescence level, and distance of object from camera and incidence angle of taking images introduces several foreign characteristics in images, making diagnostic process a difficult and cumbersome task. Image enhancement can play a vital role in this regard. This paper presents a Homomorphic filtering based framework for processing digital images of new born baby for detection of bilirubin level. The first basic step in this approach, we have used bilirubin strips images to detect the level using cameras and processed these images using Homomorphic filtering to avoid affect of luminescence. Results are evaluated statistically by finding correlation of actual bilirubin level with original image and filtered image. The higher is the correlation in filtered image toward unity as compared to unfiltered image, the better is the work performed by Homomorphic filtering.