In this letter, we present a data-driven method for scene parsing of road scenes to utilize single-channel near-infrared (NIR) images. To overcome the lack of data problem in non-RGB spectrum, we define a new color space and decompose the task of deep scene parsing into two subtasks wit...
RGBNIR/S-Tracking-Results-Datasets-and-Methods 👉 : Our contributions to the MMVOT community UniMod1K: Towards a More Universal Large-Scale Dataset and Benchmark for Multi-modal Learning. Xue-Feng Zhu, Tianyang Xu, Zongtao Liu, Zhangyong Tang, Xiao-Jun Wu, and Josef Kittler. IJCV 2024....
Pixel-aligned RGB-NIR Stereo Imaging and Dataset for Robot VisionJinnyeong Kim Seung-Hwan BaekPOSTECHAbstractIntegrating RGB and NIR stereo imaging provides com-plementary spectral information, potentially enhancingrobotic 3D vision in challenging lighting conditions. How-ever, existing datasets and imaging...
Scene recognition and classification is an important and challenging branch of computer vision. Also, image fusion is a very well-known method in image processing. In this paper, fusion of RGB and NIR images is applied to improve the performance of scene classification. The proposed fusion ...
Alternatively, instead of multiple RGB images, a “range image” can be used to get 3-D information about the objects in a scene. Unlike a regular image in which each pixel contains light intensity information, a pixel of a range image contains depth information. An easy and inexpensive meth...
A collection of deep learning based RGB-T-Fusion methods, codes, and datasets. The main directions involved are Multispectral Pedestrian Detection, RGB-T Aerial Object Detection, RGB-T Semantic Segmentation, RGB-T Crowd Counting, RGB-T Fusion Tracking. -
The method was evaluated through three datasets: the hyperspectral dataset, the RGB image dataset, and the fused dataset. Each dataset consists of four classes of VS with varying freshness levels. The datasets contain 462 bands of hyperspectral data, which was reprocessed into a single channel of...
Currently, there exist several remote sensing datasets derived from satellite and aerial imagery ready for training DL models for LULC mapping (Table1). However, they still suffer from some limitations, particularly the following factors that complicate their application with DL models: (1)First, non...
On numerous well-known large-scale datasets, existing deep convolutional neural networks have produced excellent results. However, these networks also have a number of drawbacks that are very inconvenient for practical applications, including a large model size and slow operation caused by the ...
Salient object detection, which simulates human visual perception in locating the most significant object(s) in a scene, has been widely applied to various