Beyond Color Correction : Skin Color Estimation In The Wild Through Deep Learning

Clouet, Axel; Viola, Célia; Vaillant, Jacques

doi:10.2352/issn.2470-1173.2020.5.maap-082

Cited by 6 publications

(8 citation statements)

References 6 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…However, certain prior studies [5,7] predominantly featured images under diverse lighting conditions, with limited variation in skin tones. This consideration prompts the authors of this project to explore data collection methods, drawing inspiration from [20] for image acquisition as a foundational framework.…”

Section: Data Pre-processing Methodsmentioning

confidence: 99%

“…Kips et al [20] developed a skin tone classifier for online makeup product shopping. Sobham et al [4] developed a method to assess the progress of wound healing, which has had a substantial impact on treatment decisions.…”

Section: Related Workmentioning

confidence: 99%

“…The dataset's magnitude must guarantee the effective training and generalisation of the neural network model, with a minimum benchmark of 2750 images per specific head rotation. Considering image capture across 15 lighting conditions for each head rotation, a minimum of 183 individuals must participate in data collection, emphasising that a larger dataset size confers advantages for enhancing model performance [20,26]. Furthermore, to facilitate testing and validation with colour-corrected image data, all captured images must incorporate a reference colour chart, crucial for computing colour histograms and applying colour correction to the images if deemed necessary.…”

Section: Data Gathering and Dataset Creation 211 Output Dataset Requi...mentioning

confidence: 99%

“…Table 1 outlines the hyperparameters for model training, initially set based on [20], with values tuned for optimal accuracy through a manual random search method.…”

Section: Hyperparameter Tuningmentioning

confidence: 99%

See 3 more Smart Citations

Skin Tone Estimation under Diverse Lighting Conditions

Mbatha,

Booysen,

Theart

2024

J. Imaging

View full text Add to dashboard Cite

Knowledge of a person’s level of skin pigmentation, or so-called “skin tone”, has proven to be an important building block in improving the performance and fairness of various applications that rely on computer vision. These include medical diagnosis of skin conditions, cosmetic and skincare support, and face recognition, especially for darker skin tones. However, the perception of skin tone, whether by the human eye or by an optoelectronic sensor, uses the reflection of light from the skin. The source of this light, or illumination, affects the skin tone that is perceived. This study aims to refine and assess a convolutional neural network-based skin tone estimation model that provides consistent accuracy across different skin tones under various lighting conditions. The 10-point Monk Skin Tone Scale was used to represent the skin tone spectrum. A dataset of 21,375 images was captured from volunteers across the pigmentation spectrum. Experimental results show that a regression model outperforms other models, with an estimated-to-target distance of 0.5. Using a threshold estimated-to-target skin tone distance of 2 for all lights results in average accuracy values of 85.45% and 97.16%. With the Monk Skin Tone Scale segmented into three groups, the lighter exhibits strong accuracy, the middle displays lower accuracy, and the dark falls between the two. The overall skin tone estimation achieves average error distances in the LAB space of 16.40±20.62.

show abstract

Section: Data Pre-processing Methodsmentioning

confidence: 99%

Section: Related Workmentioning

confidence: 99%

Section: Data Gathering and Dataset Creation 211 Output Dataset Requi...mentioning

confidence: 99%

“…Table 1 outlines the hyperparameters for model training, initially set based on [20], with values tuned for optimal accuracy through a manual random search method.…”

Section: Hyperparameter Tuningmentioning

confidence: 99%

See 2 more Smart Citations

Skin Tone Estimation under Diverse Lighting Conditions

Mbatha,

Booysen,

Theart

2024

J. Imaging

View full text Add to dashboard Cite

show abstract

“…As a color regression loss, we propose to use mse − lab , the mean squared error in the CIE L * a * b * space. Introduced for neural networks in [22], the mse − lab loss inherits from the perceptual properties of the color distance CIE ∆E* 1976 [28] which is key for color estimation problems. The color regression loss is described in Equations 2 and 3 for the discriminator and the generator respectively, where D color is the makeup color regression output of the discriminator, and c xi i = C m (x i ) the color label for image x i obtained using our weak model:…”

Section: Color Regression Lossmentioning

confidence: 99%

CA-GAN: Weakly Supervised Color Aware GAN for Controllable Makeup Transfer

Kips,

Gori,

Perrot

et al. 2020

Preprint

Self Cite

View full text Add to dashboard Cite

While existing makeup style transfer models perform an image synthesis whose results cannot be explicitly controlled, the ability to modify makeup color continuously is a desirable property for virtual try-on applications. We propose a new formulation for the makeup style transfer task, with the objective to learn a color controllable makeup style synthesis. We introduce CA-GAN, a generative model that learns to modify the color of specific objects (e.g. lips or eyes) in the image to an arbitrary target color while preserving background. Since color labels are rare and costly to acquire, our method leverages weakly supervised learning for conditional GANs. This enables to learn a controllable synthesis of complex objects, and only requires a weak proxy of the image attribute that we desire to modify. Finally, we present for the first time a quantitative analysis of makeup style transfer and color control performance.

show abstract