2k images

KAIST Computer Vision Project

shivom bhargava

Updated 2 years ago

4.3k

views

175

downloads

Maintainer

shivom bhargava

Metrics

mAP is equal to the average of the Average Precision metric across all classes in a model. Learn more

mAP

57.9%

Precision measures how often your model's predictions are correct. Learn more

Precision

40.9%

Recall measures what percentage of relevant labels were successfully identified. Learn more

Recall

61.5%

Try This Model

Drop an image or

browse your device

Description

KAIST dataset is originally based on Thermal Infrared and corresponding RGB image pairs. I have made a model that can convert night-time infrared to day-time RGB using TIC-cGAN (https://arxiv.org/pdf/1810.05399). I have trained the YOLOv5 model on daytime RGB images and tested on generated fake day-time images from night-time infrared input to check mAP of proposed approach.

Use This Trained Model

Try it in your browser, or deploy via our Hosted Inference API and other deployment methods.

Try In Browser Deploy Model

Cite This Project

LICENSE

Public Domain

If you use this dataset in a research paper, please cite it using the following BibTeX:

                        
        highlighted
        
            
        
     
@misc{
                            kaist-xbctd_dataset,
                            title = { KAIST Dataset },
                            type = { Open Source Dataset },
                            author = { shivom bhargava },
                            howpublished = { \url{ https://universe.roboflow.com/shivom-bhargava/kaist-xbctd } },
                            url = { https://universe.roboflow.com/shivom-bhargava/kaist-xbctd },
                            journal = { Roboflow Universe },
                            publisher = { Roboflow },
                            year = { 2022 },
                            month = { oct },
                            note = { visited on 2025-03-24 },
                            }