VivaKitAI Assistive Vision & Real-Time Directional Voice Guidance
Real-time edge computer vision and directional audio mobility assistant for the visually impaired. Integrates lightweight SSDLite-MobileNetV3 deep learning object detection with monocular geometric optical depth regression to localize navigation obstacles (pedestrians, furniture, vehicles, doors) and estimate approximate distance in meters from pinhole camera geometry. Partitions the visual field into functional left, center walking path, and right sectors, generating clear offline spoken directives (e.g., "Warning: Chair directly ahead, 1.1 meters. Step right.") with intelligent repetition suppression and top-down radar mapping without cloud latency or viral copyleft licences.
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
| Video Processing Throughput | 24.0 FPS (41.8 ms latency on standard CPU) |
| Directional Sector Accuracy (synthetic) | 100.0% (Synthetic hallway navigation suite - not real camera footage) |
| Distance Estimates | Pinhole geometry with class-height priors - approximate, for guidance only |
| Speech Synthesis Latency | 93.7 ms (Offline zero-cloud local TTS) |
Evaluated on multi-obstacle hallway and pathway navigation benchmarks. Strictly adheres to permissive BSD-3 and MPL-2.0 licensing with zero AGPL copyleft encumbrance. Includes automated pytest smoke test suite, interactive Gradio 6 console, 2D top-down radar view, and 15 viva voce questions with optical geometry.
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
Microsoft COCO Dataset (CC BY 4.0, cocodataset.org) evaluated on synthetic multi-obstacle indoor and outdoor navigation benchmarks by VivaKit.
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.