Brightening and denoising lowlight images

Yang, Xinwei; Lin, Xiangbo

doi:10.1109/icics.2015.7459919

Cited by 2 publications

(6 citation statements)

References 23 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Moreover, notice that the denoising quality is in fact quite good for our method (compare the text in the boxed regions). Nonlocal means denoising of a low-light image using [43] and the proposed method, with and without infrared data [33]. Notice that we obtain better denoising by taking the infrared information into account (see zoomed sections).…”

Section: Hyperspectral Denoisingmentioning

confidence: 83%

“…14. Nonlocal means denoising of a low-light image using [43] and the proposed method, with and without infrared data [33]. Notice that we obtain better denoising by taking the infrared information into account (see zoomed sections).…”

Section: Hyperspectral Denoisingmentioning

confidence: 83%

“…Moreover, notice that the denoising quality is in fact quite good for our method (compare the text in the boxed regions). (e) [43].…”

Section: Hyperspectral Denoisingmentioning

confidence: 99%

“…Finally, we use our fast algorithm for NLM denoising of low-light images using additional infrared data [33]. In Figure 14, we have shown a visual result which compares our method, both with and without infrared data, and [43] (an optimization method). We set the patch and search window sizes as 7×7 and 17 × 17.…”

Section: E Low-light Denoisingmentioning

confidence: 99%

See 3 more Smart Citations

Fast High-Dimensional Bilateral and Nonlocal Means Filtering

Nair

Chaudhury

2019

IEEE Trans. on Image Process.

View full text Add to dashboard Cite

Existing fast algorithms for bilateral and nonlocal means filtering mostly work with grayscale images. They cannot easily be extended to high-dimensional data such as color and hyperspectral images, patch-based data, flow-fields, etc. In this paper, we propose a fast algorithm for high-dimensional bilateral and nonlocal means filtering. Unlike existing approaches, where the focus is on approximating the data (using quantization) or the filter kernel (via analytic expansions), we locally approximate the kernel using weighted and shifted copies of a Gaussian, where the weights and shifts are inferred from the data. The algorithm emerging from the proposed approximation essentially involves clustering and fast convolutions, and is easy to implement. Moreover, a variant of our algorithm comes with a guarantee (bound) on the approximation error, which is not enjoyed by existing algorithms. We present some results for high-dimensional bilateral and nonlocal means filtering to demonstrate the speed and accuracy of our proposal. Moreover, we also show that our algorithm can outperform state-of-the-art fast approximations in terms of accuracy and timing.

show abstract

Section: Hyperspectral Denoisingmentioning

confidence: 83%

Section: Hyperspectral Denoisingmentioning

confidence: 83%

“…Moreover, notice that the denoising quality is in fact quite good for our method (compare the text in the boxed regions). (e) [43].…”

Section: Hyperspectral Denoisingmentioning

confidence: 99%

Section: E Low-light Denoisingmentioning

confidence: 99%

See 2 more Smart Citations

Fast High-Dimensional Bilateral and Nonlocal Means Filtering

Nair

Chaudhury

2019

IEEE Trans. on Image Process.

View full text Add to dashboard Cite

show abstract

“…Restored Backlit Image. (a) Original image; (b) Yelmanov's model [23]; (c) Leow's model [24]; (d) Xin's model[20]; (e) the proposed optimal stretched image.…”

mentioning

confidence: 99%

Single Backlit Image Enhancement

2020

View full text Add to dashboard Cite

Backlit images are from an excessive reflection of light being opposite to a capturing device. The existing image enhancement methods cannot be directly applied to the backlit images because they are not designed to enhance both broad light and dark regions simultaneously. Moreover, the techniques have several limitations about over-saturation or losing contrast. This paper presents a single backlit image enhancement method based on the novel full-piecewise non-linear automatic stretching, without input parameter provided by a user, e.g., gamma and so on. The computer simulation results confirm that the proposed approach can (i) undoubtedly reveal hidden details in the dark region; (ii) preserve features and color of common and light (over-brightness) regions, and (iii) increase a local contrast of dark areas. The proposed approach tested on Li's backlit image database and several backlit images from commercial devices. The simulation results demonstrate the efficiency of the proposed approach and its advantages over the cutting-edge backlit image enhancement methods in perceptual quality.

show abstract

Brightening and denoising lowlight images

Cited by 2 publications

References 23 publications

Fast High-Dimensional Bilateral and Nonlocal Means Filtering

Fast High-Dimensional Bilateral and Nonlocal Means Filtering

Single Backlit Image Enhancement

Contact Info

Product

Resources

About