Statistical Learning Approach for the Detection of Abnormalities in Cancer Cells for Finding Indication of Metastasis
Rajdeep Chakraborty, Anupam Ghosh, Jyotsna Kumar Mandal, Tanupriya Choudhury, Prasenjit Chatterjee · 2024
Statistical learning is one of the advanced techniques under operations research to analyze the post-image processing techniques over the cancerous image to understand the status of image assessment parameters. This chapter introduces the invariant shape descriptor tool with geodesic transformation, as well as z-transformation of carcinoma images. Cancer is the most unpredictable disease in which the prediction of healing directions changes abruptly. Cancer patients first go for an X-ray, computed tomography scan or positron emission computed tomography scan. The chapter shows the edge computation approaches over carcinoma images. It lists various image assessment parameters: true positive, true negative, false negative, accuracy, precision, recall/sensitivity, miss rate, specificity, prevalence, F1 score, critical success index, positive likelihood ratio, negative likelihood ratio, and diagnostic odds ratio. The chapter shows the circularity values of the different forms of shape and provides an idea of how basic shapes are present within the target image.