VivaKitIndian Traffic Sign Recognition System
Embedded-friendly computer vision system recognizing Indian Road Congress (IRC: 67-2012) regulatory, cautionary, and informatory traffic signs with Motor Vehicles Act statutory interpretations and Grad-CAM visual explanations. Built with PyTorch and Torchvision under a 100% commercially compliant license (zero copyleft restrictions). Evaluated on 222 held-out real road camera test images, delivering 86.04% Top-1 accuracy, 91.44% Top-3 accuracy, and 84.53% Macro F1 at ~8 ms/image on a laptop GPU with a compact 4.45 MB footprint. One of the 15 classes (Cross Road) is a disclosed weak spot the project can analyse and improve - ideal viva discussion material.
Secure checkout by Dodo Payments (UPI, cards). Your personalised download link is e-mailed within 24 hours. One purchase covers one project team of up to 4 students. See terms and refunds.
Our honest results
| Top-1 Classification Accuracy | 86.04% (Held-out 222 road camera test images) |
| Top-3 Candidate Accuracy | 91.44% (Ideal for automotive ADAS temporal fusion) |
| Macro F1 Score | 84.53% (14 of 15 classes at 75-100% F1; CROSS_ROAD weak at 0% - disclosed) |
| Inference Latency | ~8 ms/image on RTX laptop GPU |
| Model Footprint | 4.45 MB (Optimized for embedded microcontrollers & dashcams) |
Evaluated strictly on 222 held-out real Indian road camera test images (CC BY 4.0). Includes 4.45 MB trained PyTorch weights, 15-class IRC dataset, Gradio 6 console with live Grad-CAM visual attention heatmaps and Motor Vehicles Act legal rules, automated pytest suite, and 15 viva voce questions on IRC:67-2012, depthwise separable convolutions, Hard-Swish, and Grad-CAM mathematics.
What's included
- Readable source codeSmall, commented modules you can explain line by line in your viva.
- Demo appA local web app to show your examiners - runs offline on your laptop.
- Trained model + reference resultsSo you can compare your own runs fairly.
- Learning guide with viva Q&ABackground, design choices explained, limitations and extension ideas.
- One-click setup & testsWindows setup script and tests that prove it works on your PC.
- Dataset download scriptFetches the public data from its original source, with the citation.
Dataset credit
Indian Traffic Sign Classification Dataset by Kannan Wisen (CC BY 4.0). Indian Road Congress Standards (IRC: 67-2012).
Academic integrity
We sell code and teaching, not finished projects. You must understand, extend and write up the work yourself, acknowledge the kit and cite the dataset. No ready-made reports.