Developed a traffic sign recognition system using TensorFlow to classify and recognize traffic signs from images. Utilized
the German Traffic Sign Recognition Benchmark (GTSRB) dataset, implementing data augmentation for enhanced
performance. Used two models for comparison:
Model 1: EfficientNetB0 (pre-trained) with additional dense layers, achieving 73% accuracy.
Model 2: Custom CNN with two 5x5 kernel layers, two 3x3 kernel layers, and one dense layer, achieving nearly 99%
accuracy.