Quality Control Model for Electrospun Nanofibers through Image Analysis

Electrospinning Nanofibers are extensively used in progressive fields such as biomedical engineering for tissue scaffolds, filtration for air and water, and energy storage, among others, due to their high surface area-to-volume ratio. Nevertheless, one of the most frequent problems in this area is the inability to exercise strict control over the quality of a particular production batch, which occasionally results in a stark fluctuation in performance. This study addresses this issue by proposing the Efficient Slime Mould Algorithm fine-tuned Adaptive Deep Residual Network (ESMA-ADRN), designed to improve the quality valuation of electrospun nanofibers over advanced image examination. The dataset employed in this research includes images of electrospun nanofiber images, which are subjected to preprocessing through a median filter as a denoising technique. The process of feature extraction has been carried out using Principle Component Analysis (PCA) to determine the most useful feature space for classification. The results of the proposed ESMA-ADRN models show notable numeric values when compared to other models that lead to high achievements, such as accuracy maximum of 94.30 %, precision 96.58 %, sensitivity 93.04 %, specificity 93.72 %, and F-score of 94.77%.Future work should continue to compile more scenarios for the trained model to cover more possibilities and the adjustment and refining of the model parameter for better performance in many manufacturing situations.