Category Ranking
98%
Total Visits
921
Avg Visit Duration
2 minutes
Citations
20
Article Abstract
Although huge progress has been made on scene analysis in recent years, most existing works assume the input images to be in day-time with good lighting conditions. In this work, we aim to address the night-time scene parsing (NTSP) problem, which has two main challenges: 1) labeled night-time data are scarce, and 2) over- and under-exposures may co-occur in the input night-time images and are not explicitly modeled in existing pipelines. To tackle the scarcity of night-time data, we collect a novel labeled dataset, named NightCity, of 4,297 real night-time images with ground truth pixel-level semantic annotations. To our knowledge, NightCity is the largest dataset for NTSP. In addition, we also propose an exposure-aware framework to address the NTSP problem through augmenting the segmentation process with explicitly learned exposure features. Extensive experiments show that training on NightCity can significantly improve NTSP performances and that our exposure-aware model outperforms the state-of-the-art methods, yielding top performances on our dataset as well as existing datasets.
Download full-text PDF |
Source |
|---|---|
| http://dx.doi.org/10.1109/TIP.2021.3122004 | DOI Listing |
Publication Analysis
Top Keywords
Similar Publications
Accurate recognition of UAVs on multi-scenario perception with YOLOv9-CAG.
Sci Rep
July 2025
Information Center, Yunnan Medical Health Vocational College, Kunming, 650033, China.
As the application of Unmanned Aerial Vehicles(UAVs) becomes increasingly widespread, the identification of UAVs is of great significance in the field of security. The research of advanced identification technology can effectively deal with the illegal invasion of UAVs and reduce the threat to aviation safety. However, during the recognition process, the effectiveness of UAVs identification is often compromised in long-distance and complex environments, particularly in night-time scenarios, where accurately and reliably identifying UAVs remains a significant challenge.
View Article and Find Full Text PDFMonocular blur impairs heading judgements from optic flow.
Iperception
February 2025
National Centre for Optics, Vision and Eye Care, Faculty of Health and Social Sciences, University of South-Eastern Norway, Kongsberg, Norway.
Monocular blur sometimes impairs locomotion; however, it is not always clear when this will happen. Optic flow (the apparent motion of scene texture elements that occurs during self-motion) provides powerful signals about the direction of travel. Here, we test whether monocular blur impairs heading perception from optic flow compared to full vision under various levels of optic flow degradation.
View Article and Find Full Text PDFA hyperspectral open-source imager (HOSI).
BMC Biol
January 2025
Centre for Ecology & Conservation, University of Exeter, Penryn, UK.
Background: The spatial and spectral properties of the light environment underpin many aspects of animal behaviour, ecology and evolution, and quantifying this information is crucial in fields ranging from optical physics, agriculture/plant sciences, human psychophysics, food science, architecture and materials sciences. The escalating threat of artificial light at night (ALAN) presents unique challenges for measuring the visual impact of light pollution, requiring measurement at low light levels across the human-visible and ultraviolet ranges, across all viewing angles, and often with high within-scene contrast.
Results: Here, I present a hyperspectral open-source imager (HOSI), an innovative and low-cost solution for collecting full-field hyperspectral data.
SRNSD: Structure-Regularized Night-Time Self-Supervised Monocular Depth Estimation for Outdoor Scenes.
IEEE Trans Image Process
October 2024
Deep CNNs have achieved impressive improvements for night-time self-supervised depth estimation form a monocular image. However, the performance degrades considerably compared to day-time depth estimation due to significant domain gaps, low visibility, and varying illuminations between day and night images. To address these challenges, we propose a novel night-time self-supervised monocular depth estimation framework with structure regularization, i.
View Article and Find Full Text PDFNight-time scene parsing aims to extract pixel-level semantic information in night images, aiding downstream tasks in understanding scene object distribution. Due to limited labeled night image datasets, unsupervised domain adaptation (UDA) has become the predominant method for studying night scenes. UDA typically relies on paired day-night image pairs to guide adaptation, but this approach hampers dataset construction and restricts generalization across night scenes in different datasets.
View Article and Find Full Text PDF