Robust Magnification Independent Colon Biopsy Grading System over Multiple Data Sources

Automated grading of colon biopsy images across all magnifications is challenging because tailored segmentation and dependent features on each magnification. This work presents a novel approach robust magnification-independent cancer framework to distinguish into four classes: normal, well, moderate, poor. The contribution this research develop magnification invariant hybrid feature set comprising cartoon feature, Gabor wavelet, wavelet moments, HSV histogram, color auto-correlogram, morphological that can be used characterize different grades. Besides, the classifier modeled as multiclass structure with six binary class Bayesian optimized random forest (BO-RF) classifiers. study uses datasets (two collected from Indian hospitals—Ishita Pathology Center (IPC) 4X, 10X, 40X Aster Medcity (AMC) 20X, 40X—two benchmark datasets—gland (GlaS) 20X IMEDIATREAT 10X) multiple microscope magnifications. Experimental results demonstrate proposed method outperforms other methods for in terms accuracy (97.25%-IPC, 94.40%-AMC, 97.58%-GlaS, 99.16%-Imediatreat), sensitivity (0.9725-IPC, 0.9440-AMC, 0.9807-GlaS, 0.9923-Imediatreat), specificity (0.9908-IPC, 0.9813-AMC, 0.9907-GlaS, 0.9971-Imediatreat) F-score 0.9441-AMC, 0.9780-GlaS, 0.9923-Imediatreat). generalizability model any magnified input image validated by training one dataset testing another dataset, highlighting strong concordance classification evidencing its effective use first level automatic second opinion.