Detection Of Malignant Cancerous Nuclei Using Quantum Hadamard Edge Detection Algorithm

Abstract

Cancer is a hazardous ailment, becoming even more ominous due to its capacity to remain unnoticed. The significance of cancer research lies in its role in enhancing both the detection and treatment of this disease, ultimately leading to improved quality of life for affected individuals. The process of edge detection assumes a pivotal role in image processing, as it aids in the identification of object boundaries within images. In recent years, quantum computing has garnered substantial attention for its potential to address complex problems more efficiently than classical computers. In this work, we employed edge-detection technique to annotate images related to breast cancer. We have attempted the portion of cancerous nuclei in one biopsy data sample with several malignant nuclei and validated these findings through manual examination. We have used Quantum Hadamard Edge Detection Algorithm (QHED) to identify the nuclei edges. We implemented the algorithm on quantum simulator platforms ie qiskit and Amazon Web Services (AWS). We present the results from implementation on 4 qubits in AWS simulators i.e. State Vector (SV) Simulator and Density Matrix (DM) Simulator. In qiskit implementation, the algorithm was implemented on 4096 x 4096 pixels of biopsy image, that required 24 qubits in the qiskit platform when performing on a CPU. Further, the research presented herein compares the outcomes of Canny edge detection and the results achieved through quantum edge detection from real data and we observe that the results from the quantum approach are at par with the results obtained from the classical approach.