Data Set_Smartphone Screen Damage Detection.zip

This project aims to automate the detection of cracks in smartphone screens to ensure fairness and standardization in the used phone market. It employs a Convolutional Neural Network (CNN) model trained on a custom image dataset to classify screens as cracked or non-cracked, achieving an accuracy of 92%. Unlike traditional manual inspection methods, the CNN-based approach offers consistent, objective, and scalable evaluation. Challenges include detecting rare crack patterns, handling variable lighting conditions, and mitigating overfitting due to limited data. Future improvements will explore advanced architectures and data augmentation techniques to enhance model robustness and generalizability.