The Xilinx NEXUS is a novel initiative of SRMIST, made possible by grants from Xilinx, Inc. It aims to support Women in Technology (WIT) programs within the university, by providing the necessary funding.
Under the XilinxNexus program, which was initiated as a Xilinx WIT awardee for 2021, we have begun developing and porting an ML model to Xilinx hardware.
Breast cancer is a prevalent disease among Indian women, with poor prognosis at an early stage. Unfortunately, most cancer cases are detected only at an advanced stage due to a shortage of radiologists, pathologists, and trained technicians in our country. Additionally, the diagnosis procedure is operator-dependent and requires experienced pathologists, which can lead to misdetection due to human factors such as exhaustion and insufficient concentration. In response to this need, we are developing a computer-assisted diagnostic system based on recent developments in Deep Learning (DL) methods/networks, namely capsule networks and transfer learning.
This project aims to provide an effective screening and diagnostic system for the early detection of breast cancer, using DL methods to analyze digital mammograms and histopathology images. The Xilinx embedded hardware developer kit will interface with both the digital mammogram unit and the camera unit of the microscope, which is used to investigate tissue specimens. The DL model will provide screening and diagnosis reports for images acquired through the DM unit, providing predictions for Benign (B) and Malignant (M) findings.
Research
Early detection of breast cancer.
Under the XilinxNexus program, which was initiated as a Xilinx WIT awardee for 2021, we have begun developing and porting an ML model to Xilinx hardware.
Breast cancer is a prevalent disease among Indian women, with poor prognosis at an early stage. Unfortunately, most cancer cases are detected only at an advanced stage due to a shortage of radiologists, pathologists, and trained technicians in our country. Additionally, the diagnosis procedure is operator-dependent and requires experienced pathologists, which can lead to misdetection due to human factors such as exhaustion and insufficient concentration. In response to this need, we are developing a computer-assisted diagnostic system based on recent developments in Deep Learning (DL) methods/networks, namely capsule networks and transfer learning.
This project aims to provide an effective screening and diagnostic system for the early detection of breast cancer, using DL methods to analyze digital mammograms and histopathology images. The Xilinx embedded hardware developer kit will interface with both the digital mammogram unit and the camera unit of the microscope, which is used to investigate tissue specimens. The DL model will provide screening and diagnosis reports for images acquired through the DM unit, providing predictions for Benign (B) and Malignant (M) findings.