Renal Pathology Images Segmentation Based on Improved Cuckoo Search with Diffusion Mechanism and Adaptive Beta-Hill...

Area of Science:

• Nephrology
• Medical Imaging
• Computational Intelligence

Background:

• Lupus Nephritis (LN) significantly increases morbidity and mortality in systemic lupus erythematosus.
• Nephropathology remains the gold standard for LN diagnosis, requiring expert evaluation of histopathological images.
• Accurate segmentation of renal pathological images is crucial for effective LN diagnosis and management.

Purpose of the Study:

• To develop an advanced image segmentation technique for Lupus Nephritis (LN) histopathological images.
• To improve the accuracy and efficiency of pathological image analysis in diagnosing LN.
• To introduce a novel optimization algorithm for multi-threshold image segmentation.

Main Methods:

• A 2D Rényi entropy multi-threshold image segmentation method was developed.
• The method utilizes an improved Cuckoo Search (CS) algorithm, incorporating a Diffusion Mechanism (DM) and Adaptive β-Hill Climbing (AβHC) strategy, termed the DMCS algorithm.
• The DMCS algorithm was validated on 30 benchmark functions (IEEE CEC2017) and applied to renal pathological images.

Main Results:

• The DMCS algorithm demonstrated enhanced capability in finding optimal solutions due to the integration of DM and AβHC strategies.
• The proposed multi-threshold image segmentation method achieved strong performance in segmenting renal pathological images, as indicated by PSNR, FSIM, and SSIM metrics.
• Experimental results confirm the effectiveness of the DMCS-based segmentation for pathological image analysis.

Conclusions:

• The developed DMCS algorithm offers improved optimization capabilities for image segmentation tasks.
• The proposed image segmentation method is effective and beneficial for analyzing renal pathological images in the context of Lupus Nephritis.
• This research provides a valuable tool to assist pathologists in the histopathological evaluation of LN.