It was good programming, even though i had zero knowledge about this. ************************************* **Challenges Faced**: Initially, we tried using the **TensorFlow API**, but it was too complex, and the model's size became a problem. After doing more research, we decided to switch to YOLOv5s, which was simpler and worked better for our needs. Another challenge was capturing enough good-quality images for training. The model's accuracy depends on having a diverse dataset, so we ensured that each gesture was well-represented with different lighting conditions and backgrounds. We used data augmentation (rotating, blurring) to improve the model’s generalization. Finally, balancing speed and accuracy was tricky. We had to adjust settings like learning rate, batch size, number of epochs, and image sizes multiple times to get the best results.