Results 71 to 80 of about 2,765,050 (197)
Dual‐Path Wavelet Transform Image Exposure Correction Algorithm
This paper proposes an image exposure correction method, combining wavelet transforms and deep learning. It uses a dual‐path approach: luminance‐related low‐frequency components are corrected by an exposure correction network, while texture‐detail high‐frequency components are enhanced by a residual network.
Kaicheng Xu +3 more
wiley +1 more source
Using the Random Sprays Retinex Algorithm for Global Illumination Estimation [PDF]
In this paper the use of Random Sprays Retinex (RSR) algorithm for global illumination estimation is proposed and its feasibility tested. Like other algorithms based on the Retinex model, RSR also provides local illumination estimation and brightness adjustment for each pixel and it is faster than other path-wise Retinex algorithms.
Nikola Banic, Sven Loncaric
openaire +3 more sources
In this paper, we introduce a novel image dehazing algorithm based on dual‐channel prior adaptive contrast‐limited enhancement. The algorithm estimates model parameters from different perspectives based on dual‐channel prior knowledge and fuses the parameters according to the characteristics of each channel.
Chang Su +4 more
wiley +1 more source
Pre‐Trained Codebook‐Based Enhancement: A Novel Approach for Clarifying Underwater Images
This work presents a codebook‐driven enhancement network to tackle colour distortion and detail loss in underwater images. By aligning multi‐scale features with a pre‐trained VQGAN codebook and fusing shallow‐to‐deep cues, the method boosts contrast, edges and clarity without requiring large paired datasets.
Yuanxue Xin +4 more
wiley +1 more source
YOLO‐O: An Improved YOLO‐Based Framework for Vehicle Detection
This study proposes a YOLO‐O model, based on YOLOv7, for improved detection of tiny vehicles on roads. Enhancements like residual networks, Squeeze and Excitation Network (SENet), coordinate attention, and SIoU regression loss boost precision, especially for small objects.
Rabbia Mahum +4 more
wiley +1 more source
Mean‐Local Binary Pattern‐Guided Multi‐Attention Network for Low‐Light Image Enhancement
Low‐light image enhancement struggles with noise amplification, residual dark areas, artefacts and detail loss. This paper presents the MGA‐LLIEN network, which uses M‐LBP for adaptive brightness adjustment and detail recovery while reducing noise and outperforms leading methods in tests. ABSTRACT Low‐light image enhancement faces key challenges: noise
Binxin Tang +4 more
wiley +1 more source
Image Enhancement Based on Selective - Retinex Fusion Algorithm
The brightness adjustment method for the night-vision image enhancement is considered in this paper. The color RGB night-vision image is transformed into an uncorrelated color space--- the YUV space. According to the characteristics of the night-vision image, we develop the modified Retinex algorithm based on the S curve firstly, by which the luminance
Xuebo Jin 0001, Jia Bao, Jingjing Du
openaire +1 more source
LCH‐Net: A Lightweight OKLCH‐Space Decoupling Network for Archival Image Enhancement
This study puts forward the LCH‐Net framework. This method processes lightness, chroma and hue separately in the OKLCH colour space: the lightness adjustment sub‐network adaptively regulates illumination via a learnable lightness curve; the chroma adjustment sub‐network combines frequency‐domain and spatial‐domain denoising to suppress noise while ...
Liyang Yu, Ruilin Deng, Huaying Liu
wiley +1 more source
Point-based Spatial Color Sampling in Milano-Retinex : a Survey
Milano-Retinex is a family of Retinex-inspired spatial colour algorithms mainly developed for colour image enhancement. According to the Retinex theory, a Milano-Retinex algorithm takes as input an RGB image and processes the colour intensities of each ...
Michela Lecca +11 more
core +1 more source
RailNet: Railway Track Anomaly Detection via Image Processing With Hybrid Deep Learning Techniques
Our model leverages a DenseNet121 backbone combined with advanced neural network layers to achieve high accuracy in identifying and classifying track defects. The dataset, consisting of images of railway tracks with and without faults, was rigorously augmented to enhance model robustness.
Umair Saeed +8 more
wiley +1 more source

