Library management faces challenges in cataloging and retrieving books in large collections. This paper proposes a machine learning-based system that uses OCR and CNNs to automate book identification through cover, spine, and metadata recognition. By combining OCR for text extraction and CNNs for image analysis, the system improves identification accuracy, reduces manual effort, and enhances accessibility. Results demonstrate improved cataloging efficiency and consistency, providing a scalable solution for optimized library management and user experience across diverse settings.

Keywords— OCR, CNN, Library Management, Machine learning based Application.