Designed a deep learning system to classify building surface images as Cracked or Non-Cracked
Applied image preprocessing and data augmentation techniques
Implemented and compared FFNN, CNN, and LSTM architectures
Analyzed model performance using accuracy and loss metrics
