| --- |
| license: mit |
| task_categories: |
| - object-detection |
| language: |
| - en |
| tags: |
| - traffic |
| - indian-roads |
| size_categories: |
| - 1K<n<10K |
| --- |
| |
| # Indian Road Object Detection Dataset |
|
|
| ## Dataset Summary |
| This dataset contains over **7,300 high-resolution images** curated for training and evaluating object detection models in Indian road conditions. |
|
|
| - **Total Images:** 7,359 |
| - **Total Size:** ~2.7 GB |
| - **Domain:** Indian traffic (urban + highway) |
| - **Format:** JPG / PNG |
|
|
| --- |
|
|
| ## Key Features |
| Indian roads present unique challenges for computer vision: |
|
|
| - Dense and unstructured traffic |
| - Diverse vehicle types (auto-rickshaws, buses, trucks) |
| - Presence of animals (cows, goats, etc.) |
| - High scene variability and chaos |
|
|
| --- |
|
|
| ## Use Case |
| Designed for training **Transformer-based object detection models** such as RT-DETR. |
|
|
| --- |
|
|
| ## Dataset Structure |
|
|
