Morphological And Computational Profiling of Shark Dermal Denticles

Abstract

Shark dermal denticles are specialized skin structures with important morphological, hydrodynamic and biomimetic significance. Placoid scales (dermal denticles) in sharks exhibit a specialized microstructure that significantly reduces hydrodynamic drag and inhibits biofouling.  This study combines microscopic morphometric analysis with computational intelligence to characterize and classify shark dermal denticles. Skin samples of Scoliodon sp. were processed using KOH treatment and examined microscopically. A preliminary morphological analysis of a placoid scale imaged under a light microscope, revealing its characteristic ridged surface and tapered cusps. Quantitative image analysis was conducted using MATLAB and ImageJ to extract key morphometric parameters, including scale length, width, aspect ratio, and edge curvature while Support Vector Machine (SVM), Random Forest and Convolutional Neural Network (CNN) approaches were used for classification. The observed features underscore the functional adaptations of these scales for improved locomotion in aquatic environments. To enhance the analytical depth and scalability of this research, computational intelligence techniques—particularly machine learning—are being integrated into the profiling process. Denticle dimensions and shape descriptors were obtained. The denticles showed a mean length of 1.2 mm, width of 0.6 mm and aspect ratio of 2.0. The CNN achieved 92.5% accuracy, with precision, recall and F1-score of 91.8%, 93.2% and 92.5%, respectively. Grad-CAM analysis highlighted denticle margins and striated surface regions as important classification features. The findings demonstrate the potential of computational intelligence for rapid and quantitative profiling of shark dermal denticles and provide a foundation for future applications in biodiversity assessment and shark-skin-inspired biomimetic surface design.