LLVIP: A Visible-infrared Paired Dataset for Low-light Vision

By Xinyu Jia, Chuang Zhu, Minzhen Li, Wenqi Tang, Wenli Zhou. 2021 IEEE/CVF International Conference on Computer Vision Workshops (ICCVW), 2021

812 citations recorded by the source corpus.

It is very challenging for various visual tasks such as image fusion, pedestrian detection and image-to-image translation in low light conditions due to the loss of effective target areas. In this case, infrared and visible images can be used together to provide both rich detail information and effective target areas. In this paper, we present LLVIP, a visible-infrared paired dataset for low-light vision. This dataset contains 33672 images, or 16836 pairs, most of which were taken at very dark scenes, and all of the images are strictly aligned in time and space. Pedestrians in the dataset are labeled. We compare the dataset with other visible-infrared datasets and evaluate the performance of some popular visual algorithms including image fusion, pedestrian detection and image-to-image translation on the dataset. The experimental results demonstrate the complementary effect of fusion on image information, and find the deficiency of existing algorithms of the three visual tasks in very low-light conditions. We believe the LLVIP dataset will contribute to the community of computer vision by promoting image fusion, pedestrian detection and image-to-image translation in very low-light applications. The dataset is being released in https://bupt-ai-cz.github.io/LLVIP/.

Subfields: Aerospace Engineering, Computer Vision and Pattern Recognition, Media Technology

View Paper: https://www.semanticscholar.org/paper/ee1e4158395060077c12174b85abdf18ee9aaa0b

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Aerospace Engineering

In Computer Vision and Pattern Recognition

In Media Technology

Citations in the corpus

Drone-Based RGB-Infrared Cross-Modality Vehicle Detection Via Uncertainty-Aware Learning (2020, methodology): “Even 843 compared to the YOLOv5, which achieved the best publicly 844 available results on the LLVIP dataset, our method outper- 845 formed it by 5.5% and 1.7% on mAP for both modalities.”
LanguageBind: Extending Video-Language Pretraining to N-modality by Language-based Semantic Alignment (2023, background): “It aligns other modalities to images, facilitating a comprehensive understanding of various modalities such as infrared (Jia et al., 2021), depth (Kim et al., 2022), audio (Piczak, 2015), and IMU (Grauman et al., 2022).”
Visible and Infrared Image Fusion Using Deep Learning (2023, methodology): “Some VIF datasets, including TNO [176], INO [177], MFNet [178], RoadScene [61], VIFB [167], LLVIP [179] and M3FD [115], have been utilized in the VIF task as test sets to evaluate image fusion performance.”
Cross-Modality Fusion Transformer for Multispectral Object Detection (2021, background): “We compare in Table I the detection performance with and without CFT on different datasets (FLIR, LLVIP, and VEDAI).”
ImageBind One Embedding Space to Bind Them All (2023, methodology): “For the naturally available paired data, we use the (video, audio) pairs from the Audioset dataset [18], (image, depth) pairs from the SUN RGB-D dataset [67], (image, thermal) pairs from the LLVIP dataset [31] and (video, IMU) pairs from the Ego4D dataset [22].”