Analysis of facial redness by comparing VISIA and YLGTD

Abstract

Background

Erythema, characterized by redness of the skin, is a common symptom in various facial skin conditions. Recent advancements in image processing and analysis techniques have led to the development of methods for analyzing and assessing skin texture. This study aimed to investigate the correlation between the parameters of “You Look Good Today” (YLGTD) and VISIA in the detection and assessment of facial redness.

Materials and Methods

Thirty female subjects participated in this experiment, undergoing assessments using both YLGTD and VISIA. The subjects were evaluated for facial redness, and the feature count results within the red zone were measured by VISIA. YLGTD analyzed the number and percentage of red zone pixels. The assessments were conducted between [specific dates] in [location].

Results

The results demonstrated a significant positive correlation between the feature count results within the red zone measured by VISIA and the number of red zone pixels. Similarly, YLGTD exhibited a significant positive correlation with the number and percentage of red zone pixels.

Conclusion

In conclusion, our findings suggest a correlation between YLGTD and VISIA in the measurement of facial erythema. YLGTD can serve as a portable device for primary screening assessments, offering a convenient and reliable method to evaluate facial redness. This research contributes to the development of non‐invasive techniques for assessing and monitoring facial skin conditions, providing valuable insights for dermatological diagnosis and cosmetic testing.

1. BACKGROUND

Facial redness or erythema, is a common clinical symptom that can be caused by a variety of conditions, such as rosacea, contact dermatitis, acne, corticosteroid‐dependent dermatitis, and systemic lupus erythematosus. In recent years, dermatologists have observed an increase in patients experiencing erythema, often accompanied by burning, tingling, and itching sensations, due to social stress and environmental pollution. The detection and evaluation of erythema is therefore an important aspect of dermatological diagnosis and cosmetic testing. ¹ However, traditional methods for assessing erythema, such as IGA (Investigator's Global Assessment) and PSA (Psoriasis Area and Severity Index), are subjective and prone to interobserver variation. In order to improve our understanding and quantification of erythema, more accurate and objective assessment methods and techniques are needed. ² , ³

Spectral analysis is a powerful tool in the field of skin analysis and evaluation. With the advancement of spectral imaging, image processing, and optical instruments, high‐frequency ultrasound can provide detailed and accurate images of the layers of the skin, morphological features, and longitudinal depth of various skin conditions or injuries. Mexameter can measure the absorbance of the narrow‐wave spectrum of the skin, providing information on melanin and hemoglobin content. Colorimeter can accurately reflect Chroma Meter, which uses spectrophotometry to measure the L‐value of the skin to determine changes in skin color. VISIA‐CR and Antera 3D can measure skin texture, such as the depth and width of fine lines and wrinkles. This method of spectral analysis allows for the determination of melanin and hemoglobin, providing useful information on sun damage and microvascular dilation. Noninvasive devices such as these are crucial for perceiving skin erythema before it can be seen with the naked eye and provide quantitative information on severity for diagnosis and monitoring progress after various treatments. ⁴ , ⁵ , ⁶

The purpose of this study was to compare the assessment of facial erythema using two different instruments: VISIA and You look good today (YLGTD). VISIA is a highly popular and effective noninvasive measurement instrument manufactured by Canfield Scientific, Inc. It is known for its ability to capture high‐resolution images and analyze facial skin parameters simultaneously. ⁵ , ⁷ On the other hand, YLGTD is a newer mobile software developed by Hangzhou C2H4 Internet Technology Co, Ltd., China, which uses image processing techniques to analyze and evaluate user skin through smartphone‐based images. In this study, we aim to determine the correlation between the parameters analyzed by these two instruments. This will provide valuable information on the reliability of these instruments in the assessment of facial erythema.

2. METHODS

2.1. Subjective

In this study, 30 volunteers (30 females, 36.2 ± 7.8 years) were recruited from Shanghai, China. The subjects' faces were assessed for acne using the Pillsbury acne grade scale (1–2) by a laboratory specialist between March 1 and July 28, 2022. The subjects were also screened using the Submeter method. All subjects received a 28‐day free facial erythema repair mask treatment (Ruknowledge lotus nurturing pure mask, porvided by Jiangsu ZiXia BioTechnology Co, Ltd.). Only those who met the inclusion criteria and provided written informed consent were included in the study. Excluded subjects did not meet the necessary requirements for participation.

2.2. Measurements

The VISA measurements were conducted in a dedicated measurement darkroom at the laboratory of Noah Test Technology Co., Ltd. The ambient temperature and relative humidity during the measurements were 21 ± 1°C and 50% ± 10% RH, respectively. Prior to the test, participants were instructed to wash their faces with water for 30 min. Then, photographs of the front, left, and right sides of their faces were taken using the instrument. All subjects were required to complete an informed consent form before participating in the study.

2.3. Statistical analysis

To examine the correlations between VISIA and YLGTD, we utilized Pearson's correlation coefficients and Spearman's rank correlation coefficients. The Spearman's rank correlation coefficients were specifically used to determine correlations between the visual grading provided by each imaging examination. YLGTD generated correlations for each imaging analysis parameter within the red zone. All data analysis was conducted using SPSS 16.0 software (IBM Corporation).

3. RESULTS

3.1. Instrumental differences between VISIA and YLGTD

In this study, we obtained indices of the red zone using two methods: VISIA and YLGTD. Figure 1A–C depicts the results obtained from VISIA, whereas Figure 1D–F shows the results obtained from YLGTD. The comparison between the two methods is presented in Table 1.

FIGURE 1.

Images of red area took by VISIA (A–C) and by YLGTD (D–F).

TABLE 1.

Comparison of the two tools.

	VISIA	YLGTD
Operation platform	Windows	Android/iOS
Light sources	Standard/UV/cross‐polarized light	LED
Color channel	RGB	RGB
Pixels	1500 w	Phone camera
Data deposition	Local disk	Local disk/cloud
Disk space	≥60 G	None
Analyzed indices	Spots, wrinkles, textures, pores, UV spots, brown spots, red areas, porphyrins	Acne, pores, red areas, skin age, blackhead, texture, oiliness, dark circle, brown areas
Values	Feature counts, absolute scores, percentiles	Percent (area), pixels

The study utilized VISIA images and YLGTD‐based assessment images to assess the subjects over the course of the experiment. Parts (A–C) demonstrate the VISIA images taken on the first day, day 14, and day 28, respectively. Parts (D–F) show the YLGTD‐based assessment images taken on the same days. Although the VISIA images provide clearer and more detailed skin texture information, the YLGTD‐based images still display a degree of similarity. Additionally, the YLGTD, as a low‐cost app, still has the advantage of being easy to photograph, low‐cost, and easy to view at any time.

3.2. Red zone correlation

The feature counts within the red zone measured by VISIA were found to have a significant positive correlation with YLGTD. A significant positive correlation was found with the pixels and percentages measured by YLGTD on the left and right sides of the face (r = 0.480–0.531, p < 0.001) (e.g., Figure 2A,B). In contrast, relatively weak correlations were found in the middle part (r = 0.331–0.362, p < 0.05) (e.g., Figure 2C,D). In addition, no correlation was observed between data measured on the three sides (left, right, and middle) and absolute counts. No correlation was observed between the measured data of the three sides (left, right, and middle) and the absolute counts of the VISIA analysis. No correlation was observed between the pixel and percentage of specific regions collected with YLGTD (p > 0.05).

FIGURE 2.

Correlation diagrams of the values within the symmetrical red area measured by VISIA and YLGTD.

4. DISCUSSION

The skin's color is largely determined by two main chromophores: melanin in the basal layer of the epidermis and hemoglobin in the upper layer. ⁸ , ⁹ , ¹⁰ Cutaneous erythema, a reflection of inflammation, is thought to be related to the vascularity of the skin.

The absorbance of visible light at wavelengths of 420, 542, and 577 nm by hemoglobin in the dermis has been well documented in the scientific literature. Although trained physicians may be more accurate in predicting a patient's skin color, the human eye's perception and assessment of skin color can be influenced by various factors. ² , ¹¹ In a study, dermatologists' mean visual scores for skin color were similar, but internal consistency was not demonstrated using Kendall's coefficient. ¹² , ¹³ This inconsistency may be due to dermatologists' tendency to overestimate the severity of erythema and the influence of freckles or melasma on erythema identification. It has been noted that colorimetric measurements of melanin may be biased by fluctuations in skin erythema and vice versa due to partial overlap of wavelengths. As a result, techniques such as reflectance spectrophotometry have been adopted in place of visual matching of skin color with color standards for more accurate and objective assessment.

Ly et al. found a positive correlation between a* values and erythema. ¹⁴ , ¹⁵ Taylor et al. found that the CIELAB‐based color display conveys more accurate information about the two grayscale images to the viewer than an RGB‐based display. ¹⁶ This finding is significant as it suggests that the CIELAB‐based display is more effective at conveying information about the images being displayed.

One potential explanation for the inconsistent correlation coefficients within the red zone could be the difference in the color space used by the two devices. The VISIA device is an RGB‐based device that is designed to detect facial features by measuring the representative index of melanin and blood vessels under the surface of the skin. ¹⁷ On the other hand, the YLGTD device is a combination of an app and a cell phone that utilizes a single LED light source at a distance of 10–15 cm to conduct a quantitative analysis of different parts of the face in just 7 sec. Although the VISIA device may have some advantages in terms of its ability to detect facial features, the YLGTD device appears to have a greater advantage in terms of portability and convenience. It is easier to use and allows for immediate viewing of the analysis evaluation results on a mobile app. Additionally, the symmetrical distribution of the red areas on the subjects' faces, as observed in the adjacent r values in both the left and right parts of the face.

Overall, this study highlights the importance of considering the color space used in color displays and the potential impact on the accuracy of information conveyed to the viewer. Further research is needed to fully understand the relationship between the color space and the accuracy of perceived information. In addition, the difference in facial symmetry may be due to the different severity of erythema, the absolute VISIA score reflects the characteristics of the skin damage, including the size and intensity in a given parameter. In contrast, YLGTD pixels represent the number of skin lesions involved, regardless of size and intensity. Clearly, there is little correlation between the two different dimensions. However, we can barely explain the relationship between the VISIA absolute score and YLGTD. Relationship between absolute scores and YLGTD percentages, although they both emphasize the central tendency aspect of analyzing characters on the same dimension. Regarding the correlation between the objective parameters of the VISIA/YLGTD analysis between the objective parameters of the diagnosis. There was a positive correlation between them, except for the absolute scores of VISIA, which were mild and moderate, respectively. Our findings are consistent with previous reports by Wang et al. showing similar correlation coefficients between VISIA and IPP. ¹⁵

However, it is difficult for us to derive an uncorrelated relationship between VISIA and IPP, the uncorrelated relationship between visual scores and absolute VISIA scores. Overall, YLGTD showed a stronger correlation than VISIA in the assessment of visual and instrumental. This may be due to the color space applied. It has been previously reported that the RGB space itself is difficult because it does not correlate with the natural way, we perceive color. ¹⁸

The differences in instrument parameters could be due to different light sources, color models, and the resolution matched to the two instruments. In addition, we found a higher correlation for the index of central tendency, index of tendency (YLGTD percentage) rather than the index of dispersed tendency (YLGTD percentage) rather than the index of dispersed tendency (VISIA feature counts and YLGTD pixels).

A relatively strong positive linear correlation emerged between VISIA feature counts and YLGTD pixels/percentage. The reason for the difference may be due to the light resource and the different size of the analysis area of the two instruments.

5. CONCLUSION

The YLGTD and VISIA methods have correlation in regards to the number of features identified within the red zone of facial. The YLGTD APP, which utilizes a smartphone camera to acquire facial data, can serve as a practical and cost‐effective alternative for tracking skin texture measurements in everyday life.

CONFLICT OF INTEREST STATEMENT

The authors have no conflict of interest to declare.

Ma L, Huang X, Qiu Y, He Y. Analysis of facial redness by comparing VISIA and YLGTD. Skin Res Technol. 2023;29:1–5. 10.1111/srt.13356

DATA AVAILABILITY STATEMENT

The datasets used and/or analysed during the current study are available from the corresponding author on reasonable request.