AutoNaVIT : Vision-Based Path and Obstacle Segmentation Dataset for Autonomous Driving - JSON Compatible

AutoNaVIT is a curated segmentation label dataset developed to support research in autonomous navigation, scene understanding, and deep learning-based object segmentation. This release contains only the annotation labels corresponding to high-resolution frames extracted from a recorded driving sequence at Vellore Institute of Technology – Chennai Campus (VIT-C). The corresponding images will be released in Version 2 of the dataset shortly.

The dataset offers manually annotated, pixel-accurate segmentation masks for three key classes relevant to autonomous vehicle navigation:

Kerb – 1,377 instances

Obstacle – 258 instances

Path – 532 instances

All annotations were generated using Roboflow, ensuring high fidelity and consistency for real-world autonomous driving applications in urban and semi-urban environments.

Data Capture Specifications:

Source imagery was recorded using a Sony IMX890 sensor with the following specifications:

Sensor Size: 1/1.56", 50 MP

Lens: 6P, ƒ/1.8, 24mm equivalent, 1.0 µm pixels

Features: OIS, PDAF autofocus

Duration: 4 min 11 sec video

Frame Rate: 2 FPS

Total Annotated Frames: 504

Format Compatibility and Model Support:

AutoNaVIT annotations are provided in standard JSON format, enabling direct compatibility with the following 7 models:

Florence-2 OD

Paligemma

COCO

COCO – Segmentation

CreateML

COCO – MMDetection

SAM2

As the format adheres to common JSON standards, the labels can be easily adapted for use with other models or frameworks that support JSON-based annotations.

Benchmark Results:

To validate the dataset’s effectiveness, a YOLOv8-based segmentation model was trained using the full dataset (images + annotations). The model achieved:

Mean Average Precision (mAP): 96.5%

Precision: 92.2%

Recall: 94.4%

These results demonstrate the dataset’s reliability for training and evaluating segmentation models in autonomous vehicle systems.

Disclaimer and Attribution Requirement:
By accessing or using this dataset, users agree to the following terms under the Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License (CC BY-NC-ND 4.0):

Usage is permitted only for non-commercial academic and research purposes.

Proper attribution must state:
“Dataset courtesy of Vellore Institute of Technology – Chennai Campus.”
This must be included in all publications, presentations, or other dissemination formats.

Redistribution, commercial use, modification, or public hosting of the dataset in any form is prohibited without explicit written permission from VIT-C.

Use of this dataset implies acceptance of these terms. All rights not expressly granted are reserved by VIT-C.