Automatic Image Colorization Using Adversarial Training

Lal, Shamit; Garg, Vineet; Verma, Om Prakash

doi:10.1145/3163080.3163104

Cited by 4 publications

(4 citation statements)

References 12 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Such workflow can be carried out by conditional generative adversarial networks (cGAN), derived from GANs and designed to handle mapping operations [52]. Authors who used generative models as a base framework for colorization all relied on a conditional setting [37], [42]- [45], [48]. However, generative models are known for being unstable during training.…”

Section: B Deep Learning Based Colorization Techniquesmentioning

confidence: 99%

“…It is important to note that most articles on colorization include a comparison with the results obtained through the works of other authors, particularly in deep learning. However, most research teams work with public datasets (see Table I), such as ImageNet [38], [41], [45], CIFAR-10 [34], [46], and SUN [38], [40], [42], [49], which allow for comparison. At the time of writing this article, all models available online were based on parameters learned from nonaerial images.…”

Section: Evaluating the Colorizationmentioning

confidence: 99%

See 1 more Smart Citation

Deep Learning for Automatic Colorization of Legacy Grayscale Aerial Photographs

Herrault

Skupinski

2020

IEEE J. Sel. Top. Appl. Earth Observations Remote Sensing

View full text Add to dashboard Cite

Legacy grayscale aerial photographs represent one of the main available sources for studying the past state of the environment and its relationship to the present. However, these photographs lack spectral information thereby hindering their use in current remote sensing approaches that rely on spectral data for characterizing surfaces. This article proposes a conditional generative adversarial network, a deep learning model, to enrich legacy photographs by predicting color channels for an input grayscale image. The technique was used to colorize two orthophotographs (taken in 1956 and 1978) covering the entire Eurométropole de Strasbourg. To assess the model's performances, two strategies were proposed: first, colorized photographs were evaluated with metrics such as peak signal-to-noise ratio (PSNR), and structural similarity index (SSIM); second, random forest classifications were performed to extract land cover classes from grayscale and colorized photographs, respectively. The results revealed strong performances, with PSNR = 25.56 ± 2.20 and SSIM = 0.93 ± 0.06 indicating that the model successfully learned the mapping between grayscale and color photographs over a large territory. Moreover, land cover classifications performed on colorized data showed significant improvements over grayscale photographs, respectively, +6% and +17% for 1956 and 1978. Finally, the plausibility of outputs images was evaluated visually. We conclude that deep learning models are powerful tools for improving radiometric properties of old aerial grayscale photographs and land cover mapping. We also argue that the proposed approach could serve as a basis for further developments aiming to promote the use of aerial photographs archives for landscapes reconstruction.

show abstract

Section: B Deep Learning Based Colorization Techniquesmentioning

confidence: 99%

Section: Evaluating the Colorizationmentioning

confidence: 99%

Deep Learning for Automatic Colorization of Legacy Grayscale Aerial Photographs

Herrault

Skupinski

2020

IEEE J. Sel. Top. Appl. Earth Observations Remote Sensing

View full text Add to dashboard Cite

show abstract

“…Colorization is recent active research yet a difficult subject in the realm of image processing, with the goal of quickly predicting and colorizing grayscale images by analyzing image content with a computer. Existing colorizing algorithms can be classified into three categories depending on the information provided by humans: scribble-based [5][6][7][8][9][10], example-based [11][12][13][14][15], and learning-based [1][2][3][4][16][17][18][19][20][21][22][23][24][25][26][27][28][29][30][31][32] methods.…”

Section: Introductionmentioning

confidence: 99%

“…[ [28][29][30][31][32] made the prediction based on generative adversarial networks (GAN) [35] and GAN variants. Cao et al [30] replaced the U-net of the original generator with a convolutional architecture without dimensionality reduction.…”

Section: Introductionmentioning

confidence: 99%

Two-stage pyramidal convolutional neural networks for image colorization

Wei

Kuo

et al. 2021

SIP

View full text Add to dashboard Cite

The development of colorization algorithms through deep learning has become the current research trend. These algorithms colorize grayscale images automatically and quickly, but the colors produced are usually subdued and have low saturation. This research addresses this issue of existing algorithms by presenting a two-stage convolutional neural network (CNN) structure with the first and second stages being a chroma map generation network and a refinement network, respectively. To begin, we convert the color space of an image from RGB to HSV to predict its low-resolution chroma components and therefore reduce the computational complexity. Following that, the first-stage output is zoomed in and its detail is enhanced with a pyramidal CNN, resulting in a colorized image. Experiments show that, while using fewer parameters, our methodology produces results with more realistic color and higher saturation than existing methods.

show abstract

A Novel Method for Colorizing Black-And-White Videos And Images Utilising Faster R-CNN

V C,

et al. 2023

2023 International Conference on Inventive Computation Technologies (ICICT)

View full text Add to dashboard Cite

Automatic Image Colorization Using Adversarial Training

Cited by 4 publications

References 12 publications

Deep Learning for Automatic Colorization of Legacy Grayscale Aerial Photographs

Deep Learning for Automatic Colorization of Legacy Grayscale Aerial Photographs

Two-stage pyramidal convolutional neural networks for image colorization

A Novel Method for Colorizing Black-And-White Videos And Images Utilising Faster R-CNN

Contact Info

Product

Resources

About