Image Colorization Using Generative Adversarial Networks

Nazeri, Kamyar; Ng, Eric; Ebrahimi, Mehran

doi:10.1007/978-3-319-94544-6_9

Cited by 163 publications

(97 citation statements)

References 12 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Instead of generating images from random noise, conditional GAN generates an image from a given label [22] or image [23,24]. The conditional GAN is good at image-to-image translation tasks such as style transfer [25], super-resolution [26], and colorization [27]. Therefore, many studies exploited conditional GAN to synthesize aged faces.…”

Section: Related Workmentioning

confidence: 99%

Semi-Supervised FaceGAN for Face-Age Progression and Regression with Synthesized Paired Images

2020

View full text Add to dashboard Cite

The performance of existing face age progression or regression methods is often limited by the lack of sufficient data to train the model. To deal with this problem, we introduce a novel framework that exploits synthesized images to improve the performance. A conditional generative adversarial network (GAN) is first developed to generate facial images with targeted ages. The semi-supervised GAN, called SS-FaceGAN, is proposed. This approach considers synthesized images with a target age and the face images from the real data so that age and identity features can be explicitly utilized in the objective function of the network. We analyze the performance of our method over previous studies qualitatively and quantitatively. The experimental results show that the SS-FaceGAN model can produce realistic human faces in terms of both identity preservation and age preservation with the quantitative results of a decent face detection rate of 97% and similarity score of 0.30 on average.

show abstract

Section: Related Workmentioning

confidence: 99%

Semi-Supervised FaceGAN for Face-Age Progression and Regression with Synthesized Paired Images

2020

View full text Add to dashboard Cite

show abstract

“…Image colorization converts a grayscale or black and white image to a color image. 23 The prediction functions from large datasets of color images can be seen as a colorization regression problem in the continuous color space. [24][25][26] Most of the colorization algorithms differ in the methods of training data acquisition and transformation from the grayscale to color.…”

Section: Previous Studies On Color Related Applications Based On Mamentioning

confidence: 99%

Color reproduction in virtual lip makeup using a convolutional neural network

Kim

Lee

2020

Color Research & Application

View full text Add to dashboard Cite

Recently, it has become possible to examine the suitability of cosmetic products by virtual makeup techniques so that shoppers can buy products online. The virtual makeup can also be utilized at offline stores to prevent possible sanitation problems associated with swatching. Faithful color reproduction is one of the most important factors in virtual makeup applications. Thus, the color difference between the virtual and real makeup results needs to be minimized. However, most previous studies on virtual makeup focus on the recommendation of makeup style rather than on the accuracy of color reproduction. This article proposes an accurate lipstick color reproduction method based on convolutional neural network. This study indicates that the proposed method using a convolutional neural network results in the minimum value of color difference compared with linear regression and multilayer perceptron algorithms.

show abstract

“…Similarly, most image-to-image ''translating'' questions are many-to-one, such as mapping photos to edges or semantic tags, or one-to-many, such as mapping tags to realistic images. Each of these tasks has been processed by a separate algorithm [11], [12], although these methods are always the same: predicting pixel-to -pixel mapping.…”

Section: B Learning-based Approachmentioning

confidence: 99%

Image Hazing Algorithm Based on Generative Adversarial Networks

et al. 2020

View full text Add to dashboard Cite

Haze is an important factor in photography with a special aesthetic, emotional, or compositional meaning, an image hazing method is proposed based on generative adversarial network. The proposed network consists of two parts: the generator has a symmetric encoder and decoder structure with skip connection, which is used to generate hazy images; and the discriminator is a global fully convolutional network, which is used to identify the reality of the generated hazy images. The haze-free image is used as the input of the generative network to obtain a corresponding hazy image. Then the discriminative network judges the similarity between the original haze-free image and the corresponding hazy image to optimize the network parameters, which means that all the parameters of the entire network can be automatically learned through training. The loss function of the network is the combination of the GAN(Generative Adversarial Networks) loss and the REG(Regression) loss with a coefficient λ. We design a REG loss which includes feature loss loss FL and image loss loss IL. By calculating the loss, a model mapping the relationship between the two corresponding pixels from the original image and the synthetic hazy image respectively is well trained. In the last part of the paper, we have shown some impressing synthetic hazy images obtained by our model as well as analyzing the effects on images with different scenes. Moreover, we have contrasted the experimental results of our algorithm with that of other classical ones. Experiments demonstrate that the proposed algorithm has a better performance than other state-of-the-art methods on both virtual-scene and real-world images qualitatively and quantitatively. INDEX TERMS Generative adversarial networks, image hazing, regression loss, mapping from pixels to pixels.

show abstract

Image Colorization Using Generative Adversarial Networks

Cited by 163 publications

References 12 publications

Semi-Supervised FaceGAN for Face-Age Progression and Regression with Synthesized Paired Images

Semi-Supervised FaceGAN for Face-Age Progression and Regression with Synthesized Paired Images

Color reproduction in virtual lip makeup using a convolutional neural network

Image Hazing Algorithm Based on Generative Adversarial Networks

Contact Info

Product

Resources

About