Block Function Convolutional Layer Extracts spatial features from the image. Activation Function (ReLU) Adds non-linearity for better learning. Pooling Layer (Max/Average Pooling) Reduces dimensionality and keeps important information. Batch Normalization Normalizes activations, stabilizing training. Dropout Layer Prevents overfitting by randomly turning off neurons. Fully Connected (FC) Layer Learns relationships between features and outputs final scores. Output Layer Produces final predictions using Softmax/Sigmoid. Residual Blocks (ResNet) Helps train deep networks by allowing gradient flow.