Enhancement and Denoising of Near-Infrared Image with Multiscale Morphology
Zhiyong Gao, Zhifeng Ma, Xinhao Chen, Haihua Liu · 2011
Near-infrared imaging (NIR) is non-invasive and feasible for the detection of breast cancer and is widely used. But the local contrast of NIR image is low and there are many kinds of noises in the image. Therefore, enhancement and denoising are needed to facilitate diagnosis. This paper presented a method based on multiscale morphology to enhance NIR image while reduce noises. Different weights related with different objects are set to enhance meaningful dark objects such as blood vessels and shadow in NIR image. Experimental results and subjective evaluation show that the method is effective and satisfactory.