PARALLEL APPROACH FOR TUBULE GRADING IN BREAST-CANCER LESIONS

Sequential algorithms and conventional architectures for grading tubul es in digitized images of microscopic slide information of breast canc er malignancies are time-consuming and labor-intensive because of high computational complexity. This paper presents a parallel algorithm an d its corresponding VLSI architecture for grading tubules in breast ca ncer lesions in real time. The algorithm relies on a combination of cr iteria used by physicians including brightness contrast, dark edge, an d uniform density of tubule areas. The algorithm exploits concurrencie s in both the space domain and the time domain. It only requires each processing element (PE) to perform three simple operations such as add ition, subtraction, and comparison, which makes every PE simple. The e ssential parallelism in the proposed algorithm facilitates its impleme ntation using VLSI architecture, so that the time complexity is only O (N-2/k) if using kxk processing elements, where N is the dimension of the image plane. If k=N, the time complexity will be O(N), compared wi th O(N-2) using a uniprocessor. A series of experiments and simulation s have verified the high performance of the proposed algorithm and the parallel VLSI architecture.