The application of computer technology has also brought many conveniences to people. As an important part of computer technology, the application of imageprocessing technology in visual transmission system has create...
详细信息
Following the impressive development of LLMs, vision-language alignment in LLMs is actively being researched to enable multimodal reasoning and visual input/output. This direction of research is particularly relevant ...
详细信息
We propose a method for capturing high-quality images in low-light environments using multi-band near-infrared (NIR) images, which offer robustness to brightness variations and provide structural information not prese...
详细信息
The attribute-based person search task aims to find matching pedestrian images by text attributes, which is relevant in scenarios where no query image is given. However, the existing methods exhibit inferior performan...
详细信息
This work addresses the task of weakly-supervised object localization. The goal is to learn object localization using only image-level class labels, which are much easier to obtain compared to bounding box annotations...
详细信息
Recently, studies on generative models using 3D information are active. GIRAFFE, one of the latest 3D-aware generative models, shows better feature disentanglement than existing generative models because it generates ...
详细信息
images are an integral and indispensable aspect of various disciplines, such as medicine, surveillance, and the entertainment industry. However, the quality of images can be severely compromised by the presence of sen...
详细信息
The necessity of secure image transmission and storage has become more urgent in the digital era. In a variety of applications, including medical imaging, military communications, and personal data protection, image e...
详细信息
This paper proposes a chest X-ray image super-resolution reconstruction method - Deep Contrast Consistent Feature Network, which articulates the contrast consistency as a task to learn spatial contrast enhancement cur...
详细信息
ISBN:
(纸本)9783031235986;9783031235993
This paper proposes a chest X-ray image super-resolution reconstruction method - Deep Contrast Consistent Feature Network, which articulates the contrast consistency as a task to learn spatial contrast enhancement curve followed by a depth contrast network. The proposed method trains a network with visual quality measures through learning non-reference loss - Contrast Consistency Loss, which aims to overcome contrast overstretching and contrastive variation. Specifically, this network does not need to reference images at the time of loss formulation. We evaluated the proposed method for chest X-ray 5Kdatasets over several benchmarks metrics and models for quantitative and qualitative analysis. The extensive experiments report that the proposed model outperforms other approaches.
作者：
Jeong, Jong-BeomPark, Jun-HyeongLee, SoonbinRyu, Eun-Seok
Department of Computer Science Education Seoul Korea Republic of
Department of Immersive Media Engineering Seoul Korea Republic of
Multimedia Communications Group Berlin Germany
Tile-based streaming is widely adopted for viewport-Adaptive 360-degree video streaming due to its potential for bitrate reduction. However, for legacy devices equipped with a single decoder, managing multiple tile bi...
详细信息
暂无评论