Comparative analysis of classical ML models (SVM, Random Forest) and CNN architectures (LeNet5, AlexNet, VGG16, ResNet50, InceptionV3, MobileNetV2) for multi-class image classification. Includes training, evaluation, and performance benchmarking across models.