Abstract: Image classification is essential across various sectors, from healthcare to robotics. However, traditional methods often struggle with datasets containing multiple categories. In this paper, we present our innovative approach to multi-class image classification, achieving an accuracy of 86%. Our algorithm is built from scratch, tailored specifically for multi-class categorization. Through rigorous testing, we demonstrate its effectiveness, outperforming baseline methods. Our work contributes to advancing image classification, providing a robust solution for multi-class challenges.

 
Keywords: Machine learning, Binary classification, Neural Networks, Application.

| DOI: 10.17148/IARJSET.2024.11557