Hardware Implementation of Active Contour Algorithm for Fast Cancer Cells Detection

A critical process in real-time optical microscopy applications, such as cancer cells detection, is the determination of a cells position in the histopathological image. Aiming at low-cost and efficiency, a proof-of-concept study was done to distinguish between normal and abnormal cells based on optical microscopy at the LICM laboratory. A snake/active contour method was developed in which several curves move within images to find normal/abnormal cell boundaries. Abnormal cell identification typically takes more than one hour, however an implementation on FPGA technology solves this problem. A novel embedded architecture of the snake method was developed for an efficient and fast computation of active contour used in high throughput image analysis applications, where time performance is critical. This architecture allows for a scalable and a totally embedded processing on FPGA of a large number of images. The architecture of the snake method was able to detect objects from images which have irregular shapes, such as carcinoma cell types. To demonstrate the effectiveness of the approach, the architecture was implemented on Xilinx ISE 12.3-FPGA technology using VHDL structural description. It shows the possibility of using the snake method implementation on FPGA technology to detect abnormal cells in real time.