---
license: cc-by-nc-nd-4.0
task_categories:
- object-detection
- depth-estimation
- image-classification
language:
- en
size_categories:
- 1K<n<10K
---
NITCAD (National Institute of Technology Calicut Autonomous Driving) is an effort to collect and label dataset to develop and contribute to autonomous driving research in India.

India has some unique road scenarios such as
- No lanes
- New common objects on roads (like cows, autos, rickshaws)
- No clear traffic signs for pedestrains, zebra crossings etc;

So, that calls for specialized datasets to train and develop ML models for Indian roads. As a contribution, we have collected stereo and image datasets on Indian roads and labeled different classes that can be used to train Deep Learning models.

To **access** the dataset, please fill out the [Google Form](https://docs.google.com/forms/d/e/1FAIpQLScv2QXTYy3jwiAlqC9ro-lR_4UhUaWtBDNYQ4jWLZ3eUv3nSA/viewform?usp=sf_link) - https://docs.google.com/forms/d/e/1FAIpQLScv2QXTYy3jwiAlqC9ro-lR_4UhUaWtBDNYQ4jWLZ3eUv3nSA/viewform?usp=sf_link

<p align="center">
<img src="https://cdn-uploads.huggingface.co/production/uploads/632be8c900e5ff9ee1277863/cNwdlsyuDtkBMsFLiyZSM.webp" width="700">
<em>An example image from NITCAD dataset where you can observe the lack of lanes and dividers.</em>
</p>

For any questions regarding dataset/paper, feel free to email - namburisrinath@gmail.com, psankaran@nitc.ac.in with mail subject as “NITCAD Dataset — Doubts”

 **Additional Links:** 
- [Medium Blog](https://namburisrinath.medium.com/nitcad-an-object-detection-classification-and-stereo-vision-dataset-for-autonomous-navigation-f28d3fe5b7d9) post explaining more details on dataset collection.
- [Github repo](https://github.com/NamburiSrinath/NITCAD-dataset)
- [LinkedIn post](https://www.linkedin.com/posts/namburi-gnvv-satya-sai-srinath_opensource-dataset-ai-activity-6723060926322495488-AlLJ?utm_source=share&utm_medium=member_desktop)

If this work helps in your research, please consider to cite as

    @article{srinath2020nitcad,
      title={NITCAD-Developing an object detection, classification and stereo vision dataset for autonomous navigation in Indian roads},
      author={Srinath, Namburi GNVV Satya Sai and Joseph, Athul Zac and Umamaheswaran, S and Priyanka, Ch Lakshmi and Nair, Malavika and Sankaran, Praveen},
      journal={Procedia Computer Science},
      volume={171},
      pages={207--216},
      year={2020},
      publisher={Elsevier}
    }

Also you can refer to [Speed estimation using Stereo Vision images](https://ieeexplore.ieee.org/abstract/document/9031876?casa_token=qeCiQNa9m50AAAAA:lOe4ogBfc866e3gPs2s6yesqeHqJ22WElxCQxdl_luLtbeTrgb_eluUFsmMrr8040A_S8U1Lof4y) for the speed estimation by using SIFT (Scale Invariant Feature Transform), YOLO and MC-CNN.

If this work helps in your research, please consider to cite as

    @inproceedings{umamaheswaran2019stereo,
      title={Stereo Vision Based Speed Estimation for Autonomous Driving},
      author={Umamaheswaran, S and Nair, Malavika and Joseph, Athul Zac and Srinath, Namburi GNVV Satya Sai and Priyanka, Ch Lakshmi and Sankaran, Praveen},
      booktitle={2019 International Conference on Information Technology (ICIT)},
      pages={201--205},
      year={2019},
      organization={IEEE}
    }


**Note:** This project was done in 2019 and we are not actively working on this project. This dataset card is mainly created to improve visibility, searchability for the datasets via the Huggingface community.