Optimized Machine Learning for Classifying Colorectal Tissues

Ashish Tripathi*, Anuradha Misra, Kuldeep Kumar, Brijesh Kumar Chaurasia

*Corresponding author for this work

Research output: Contribution to journalArticleResearchpeer-review

9 Citations (Scopus)


Due to numerous deaths, colon cancer treatment and diagnosis are viewed as societal and financial challenges. The most severe reason for death worldwide is colorectal cancer. The classification of colon cancer tissues through images is presented in this paper as a multifaceted task. Classifying an illness at a premature stage increases its chances of existence, as late detection can be mortal which results in metastasis and a poor prognosis. The microscopic examination and classification of infected colon tissue sample images is a complex task. Also, the failure to manually detect the abnormality in the tissue by a pathologist might increase the severity of the disease. With the aid of intelligent machines, and automated diagnosis the classification of tissues from images can be done in much less time. These algorithms can learn by analyzing the patterns in the images and support the pathologist in completing the task with greater accuracy. In this research article, we proposed a tuned machine learning model, with the application of five machine learning techniques (K-Nearest Neighbor, Decision Trees, Random Forest, Categorical Boosting, and Gaussian Naive Bayes) for accurately classifying histopathological colon cancer tissues images of National Center for Tumor diseases Bank. The results demonstrate that the Categorical Boosting model has the best performance and is the most viable approach (accuracy: 0.9067, F1-Score: 0.9053, specificity: 0.9739, and sensitivity: 0.9757).
Original languageEnglish
Article number461
Pages (from-to)1-14
Number of pages14
JournalSN Computer Science
Publication statusPublished - 22 Jun 2023
Externally publishedYes


Dive into the research topics of 'Optimized Machine Learning for Classifying Colorectal Tissues'. Together they form a unique fingerprint.

Cite this