Proceedings ArticleDOI
Non-local Image Dehazing
Dana Berman,Tali Treibitz,Shai Avidan +2 more
- pp 1674-1682
TLDR
This work proposes an algorithm, linear in the size of the image, deterministic and requires no training, that performs well on a wide variety of images and is competitive with other state-of-the-art methods on the single image dehazing problem.Abstract:Â
Haze limits visibility and reduces image contrast in outdoor images. The degradation is different for every pixel and depends on the distance of the scene point from the camera. This dependency is expressed in the transmission coefficients, that control the scene attenuation and amount of haze in every pixel. Previous methods solve the single image dehazing problem using various patch-based priors. We, on the other hand, propose an algorithm based on a new, non-local prior. The algorithm relies on the assumption that colors of a haze-free image are well approximated by a few hundred distinct colors, that form tight clusters in RGB space. Our key observation is that pixels in a given cluster are often non-local, i.e., they are spread over the entire image plane and are located at different distances from the camera. In the presence of haze these varying distances translate to different transmission coefficients. Therefore, each color cluster in the clear image becomes a line in RGB space, that we term a haze-line. Using these haze-lines, our algorithm recovers both the distance map and the haze-free image. The algorithm is linear in the size of the image, deterministic and requires no training. It performs well on a wide variety of images and is competitive with other stateof-the-art methods.read more
Citations
More filters
Journal ArticleDOI
Multiscale Supervision-Guided Context Aggregation Network for Single Image Dehazing
TL;DR: Zhang et al. as mentioned in this paper proposed a multiscale supervision-guided context aggregation network (MSGCAN) based on two principles: improving feature extraction and enhancing feature mapping.
Journal ArticleDOI
Efficient and Accurate Multi-Scale Topological Network for Single Image Dehazing
TL;DR: Wang et al. as mentioned in this paper proposed a multi-scale topological network (MSTN) to fully explore the features at different scales and designed a Multi-scale Feature Fusion Module (MFFM) and an Adaptive Feature Selection Module (AFSM) to achieve the selection and fusion of features at multiple scales, so as to achieve progressive image dehazing.
Journal ArticleDOI
Single UHD Image Dehazing via Interpretable Pyramid Network
TL;DR: This work introduces the principle of infinite approximation of Taylor’s theorem with the Laplace pyramid pattern to build a model which is capable of handling 4K hazy images in real-time and proposes a Tucker reconstructionbased regularization term that acts on each branch network of the pyramid model.
Proceedings ArticleDOI
RT-VENet: A Convolutional Network for Real-time Video Enhancement
TL;DR: This work presents a novel convolutional network that can perform high-quality enhancement on 1080p videos at 45 FPS with a single CPU and performs about 10 times faster than the current real-time method on high-resolution videos.
Journal ArticleDOI
Dehazing of Satellite Images using Adaptive Black Widow Optimization-based framework
TL;DR: A novel method for de-hazing satellite imagery and outdoor camera images is proposed by modifying the transmission map used in Dark Channel Prior (DCP) method and performs better as compared with others, independent of the haze density, without losing the natural look of the scene.
References
More filters
Proceedings ArticleDOI
Visibility in bad weather from a single image
TL;DR: A cost function in the framework of Markov random fields is developed, which can be efficiently optimized by various techniques, such as graph-cuts or belief propagation, and is applicable for both color and gray images.
Journal ArticleDOI
Single image dehazing
TL;DR: Results demonstrate the new method abilities to remove the haze layer as well as provide a reliable transmission estimate which can be used for additional applications such as image refocusing and novel view synthesis.
Proceedings ArticleDOI
Fast visibility restoration from a single color or gray level image
TL;DR: A novel algorithm and variants for visibility restoration from a single image which allows visibility restoration to be applied for the first time within real-time processing applications such as sign, lane-marking and obstacle detection from an in-vehicle camera.
Proceedings ArticleDOI
Single image haze removal using dark channel prior
Kaiming He,Jian Sun,Xiaoou Tang +2 more
TL;DR: A simple but effective image prior - dark channel prior to remove haze from a single input image is proposed, based on a key observation - most local patches in haze-free outdoor images contain some pixels which have very low intensities in at least one color channel.
Journal ArticleDOI
Dehazing Using Color-Lines
TL;DR: A new method for single-image dehazing that relies on a generic regularity in natural images where pixels of small image patches typically exhibit a 1D distribution in RGB color space, known as color-lines is described.