Histopathology-driven prostate cancer identification: A VBIR approach with CLAHE and GLCM insights

• Published in: Computers in biology and medicine Vol. 182; p. 109213
• Main Authors: Rangaiah, Pramod K.B., Pradeep Kumar, B.P., Augustine, Robin
• Format: Journal Article
• Language: English
• Published: United States Elsevier Ltd 01.11.2024; Elsevier Limited
• Subjects: Algorithms; Gland classification; Gray-Level Co-Occurrence Matrix (GLCM); Histological image; Humans; Image analysis; Image contrast; Image enhancement; Image Interpretation, Computer-Assisted - methods; Image processing; Internal Medicine; Male; Medical imaging; Medical research; Other; Prostate - diagnostic imaging; Prostate - pathology; Prostate cancer; Prostatic Neoplasms - diagnostic imaging; Prostatic Neoplasms - pathology; Reclamation; Standard deviation; Support Vector Machine; Support vector machines; Visual-Based Image Reclamation (VBIR); Prostate cancer; Gland classification; Histological image; Visual-Based Image Reclamation (VBIR); Gray-Level Co-Occurrence Matrix (GLCM)
• ISSN: 0010-4825; 1879-0534; 1879-0534
• DOI: 10.1016/j.compbiomed.2024.109213

Efficient extraction and analysis of histopathological images are crucial for accurate medical diagnoses, particularly for prostate cancer. This research enhances histopathological image reclamation by integrating Visual-Based Image Reclamation (VBIR) techniques with contrast-limited adaptive Histogram Equalization (CLAHE) and the Gray-Level Co-occurrence Matrix (GLCM) algorithm. The proposed method leverages CLAHE to improve image contrast and visibility, crucial for regions with varying illumination, and employs a non-linear Support Vector Machine (SVM) to incorporate GLCM features. Our approach achieved a notable success rate of 89.6%, demonstrating significant improvement in image analysis. The average execution time for matched tissues was 41.23 s (standard deviation 36.87 s), and for unmatched tissues, 21.22 s (standard deviation 29.18 s). These results underscore the method’s efficiency and reliability in processing histopathological images. The findings from this study highlight the potential of our method to enhance image reclamation processes, paving the way for further research and advancements in medical image analysis. The superior performance of our approach signifies its capability to significantly improve histopathological image analysis, contributing to more accurate and efficient diagnostic practices. •RGB to LUV conversion and green channel extraction enhance image feature detection.•CLAHE improves image quality, aiding analysis in varied illumination conditions.•Statistical metrics offer insights into pixel distribution and image characteristics.•GLCM-based system advances image retrieval with contrast and correlation metrics.•SVM integration boosts precision in matching image features with user queries.