Automated Cancer Diagnosis using Histopathological Images

Broad domain – Medical Image Processing

Focused domain – Pre-processing, Classification of histopathological images as benign and malignant

Cancer is a foremost reason of death across the world. The most common cancers are ­of lung, liver, colorectal, stomach and breast. Histopathology is the process of microscopic examination and detailed analysis of a sample of the biopsy by a pathologist for studying the cancer growth. The input to the system will be a collection of histopathological images of tissue. The pre-processing of an image will be performed in order to enhance the quality of the image. Subsequently, the object of interest will be segmented. The useful features will be extracted and selected for further processing. After the extraction of the selected features, the suspicious regions will be classified as benign/malignant. Finally, the performance of proposed method will be compared with the other state of art methods. The outcome of the proposed work will be an automated cancer diagnosis system which would definitely be an aid for experienced pathologists by providing a second opinion for effective diagnosis in order to avoid false decisions.

No. of intern positions available in the project – 1

Principal Investigator:

Dr Deepika Koundal

Funding- Internal        Project Start Date - 01/11/2017