Analytical Use of IHC Dataset By Using Segmentation And Classification Techniques
Hasanain Hayder Razzaq, Rozaida Ghazali, Loay Edwar George · 2022 5th International Conference on Engineering Technology and its Applications (IICETA) · 2022
This paper provides the applicability of a dataset of breast cancer samples prepared in immunohistochemistry staining and applies on it common methods in the literature including image enhancement, segmentation, and classification techniques as the proof. Segmentation techniques such as watershed and highpass are tested for solving the problem of overlapping and classification methods including SVM, K-NN, and random forest are applied and the results are compared. The results show that the accuracy of RF is the best among the three methods which is 90% without segmentation methods and 93% with segmentation methods added. a precision of 95.5%. In addition the computational time for all the methods involved is acquired which shows that RF has the highest computational time but the high accuracy is worth the added extra seconds.