Lip color diversity: An intricate study

Abstract

Background

Lip investigations and characterizations in the literature are less prevalent than for skin, particularly on the topic of color diversity. However, as the consumer demand increases for a nude lip makeup result, that is, shades close to the bare lip color, the identification and modification of lip color is essential for the cosmetic industry.

Objective

The objective was to highlight lip color diversity among three ethnicities (Caucasian, African and Hispanic), through the use of a spectral color measurement device especially adapted to the lip area, and to consider lip color ethnic specificities and overlaps.

Materials and Methods

The inferior natural lip color was measured with a full‐face hyperspectral imaging system, SpectraFace (Newtone Technologies, Lyon, France), on 410 healthy women aged 19 to 68 (Caucasian French, Caucasian American, African American, and Hispanic American women). A hierarchical ascending classification, was deployed to determine clusters based on the lip colorimetric parameters along two strategies to identify the best statistical analysis to preserve the lip color diversity.

Results

Lip color is a continuous color space, with great intra‐ethnic and inter‐ethnic diversity, especially for African American women in terms of chroma and lightness. Among the two strategies of data analysis, our two‐step statistical clustering analysis yielded 11 groups (i.e., 11 lip tones), revealing an accurate representation of the scope of diversity, but also of the overlaps.

Conclusion

The 11 lip tones/colors could potentially serve as target shades for the development of a more diverse and inclusive range of lip cosmetics, such as nude lipsticks.

1. INTRODUCTION

Human lips have three fundamental characteristics: shape, color and texture. From an anatomical perspective, lips comprise four primary zones: the skin, the vermilion, the vermilion border and the oral mucosa. ¹ The transitional area between the two epithelia is located at the vermilion, the red margin of the lips. Predominantly, lip pigmentation derives from hemoglobin, with a smaller contribution from melanin. ² Several factors, including age, ³ ethnicity, ² , ⁴ hormonal status, ⁵ and external influences such as tobacco usage, ⁶ can affect lip color. Furthermore, lips may exhibit pigment anomalies, such as darker lips resulting from melanin accumulation in the epidermal basal layer, ⁷ or conditions such as vitiligo. ⁸

Skin color measurement has attracted significant attention, both due to women worldwide wishing to understand and enhance their skin tone through makeup, and to the increasing availability and precision of photo‐based color measurement tools. Yet although cosmetics for the lips is equally important as makeup for the complexion, the study of lip color is less prevalent in the literature. This gap in research is possibly due to the unique texture and size of the lips, which can render color measurement challenging.

In the realm of cosmetics, each facial feature is part of an increasingly complex ritual. Following the evolution of the foundation routine, with the addition of new steps such as sculpting, bronzing, and highlighting, lip aesthetics has now begun to undergo a similar ritualization. This progression permits the correction, enhancement, or alteration of a diverse range of lip color tones according to personal preferences. Lipstick provides women with a tool to accentuate their individuality, showcase their personality, and highlight their smile, ⁹ potentially mirroring their own self‐perception and the image they wish to project to society.

The variety of lip makeup, in terms of color range, surpasses that of skin tone makeup, and within this range, the importance of nude lip colors must not be underestimated. The relatively recent consumer demand for nude lipstick shades has encouraged the cosmetic industry to deepen its understanding of lip color diversity. In the context of makeup, a “nude” shade is intended to evoke a sense of naturalness, closely resembling one's inherent skin and/or lip tone (i.e., tone‐on‐tone). However, the complexity of “nude” arises from the fact that what one person perceives as tone‐on‐tone might differ substantially for another, given the myriad skin and lip tones and shades. A nude color is highly individualistic ‐ what one person may consider nude could be another person's red or pink. The nuances of the term “nude” explain why finding the “right” nude for each individual proves challenging. Consequently, “nude” can be conceptualized as a broad color spectrum. The precise definition of this particular color, based on the natural lip color of women across different ethnicities, forms the context of this research. Given the powerful impact of cosmetics, which extends beyond the superficial aspects of facial appearance, while also considering the crucial role of the lips within this perception, the identification and modification of lip color become essential for the cosmetic industry. Our research therefore aimed to delve deeper into lip color diversity among three ethnicities, utilizing spectral color measurement tools specially adapted for the lip area, while also considering ethnicity‐specific color variations and overlaps. The insights garnered from this study will contribute to the development of a more diverse and inclusive range of nude lipsticks.

2. METHODS

2.1. Participants

The study engaged 410 healthy women comprised of 100 Caucasian French, 107 Caucasian American, 103 African American, and 100 Hispanic American women. The age range of volunteers spanned from 19 to 68 years (mean = 42.5; standard deviation = 12.7), including 67 women aged 18–28 years, 104 women aged 29–38 years, 97 women aged 39–48 years, 86 women aged 49–58 years, and 56 women aged 59–68 years.

Selection criteria involved the use of selective lipsticks at least twice weekly. Exclusionary factors included prior facial surgical procedures and any lip tattooing for the French volunteers, and a 7% inclusion rate for cosmetic surgery among the 310 American women. The participants were advised to abstain from applying any lip products, whether moisturizer or make‐up, from the evening before their appointment.

All participants were briefed about the study objectives and provided signed informed consent, inclusive of image rights, upon their arrival. As an acknowledgment for their participation, the volunteers received a gift in France or monetary compensation in the United States.

2.2. Experimental data acquisition device

The lip color measurement was carried out using a full‐face hyperspectral imaging system, SpectraFace (Newtone Technologies, Lyon, France). The SpectraFace device is based on a high‐resolution monochrome camera (2048×2048 pixels) and a liquid crystal tunable filter (LCTF) allowing to select 30 wavelengths from 410 to 700 nm with an interval of 10nm. Two lighting pods, located on both sides, homogeneously illuminate the subject's face. The system is equipped with polarization filters in cross position to remove specular reflection and ensure relevant color analysis. The 30 individual images are captured sequentially within 3 s, thereby limiting subject movement during the acquisition. The stack of these 30 2D monochrome images acquired at different wavelengths results in a hyperspectral image, with spectral information available in each pixel. Using the spectral information available in each pixel, it is possible to compute the RGB color coordinates given a specific illuminant and observer, resulting in a standardized RGB color image. For this research, the D65 illuminant and 10° observer were used as standard conditions for color computation. The D65 illuminant corresponds to the average midday light and is often used as a reference illuminant for colorimetric measurement and analysis. The 10° observer corresponds to a standard human observer with a view angle of 10°. ¹⁰ The resulting color image is a strict conversion of the spectral reflectance without any post‐treatment or color enhancement applied, contrary to most other commercial devices. There is no specular reflection on the resulting color reconstruction, due to the cross‐polarization. This non‐contact measurement device avoids alteration of the blood flow through pressure exertion on the lips, which could potentially modify the color.

The combination of imaging technology to obtain spatial data, coupled with the analytical possibilities it provides (repositioning of ROI, computation of evenness, etc.), and hyperspectral measurement for accurate color data, make the SpectraFace an optimal supplementary tool for skin and lip spectral color characterization. This is complementary to standardized, easy‐to‐use, and deployable photo‐based color measurement devices such as the VISIA‐CR. ¹¹

2.3. Human lip color measurement

2.3.1. Measurements

The data were recorded in rooms maintaining a standardized temperature (21°C ± 3°C). In France, the study was conducted in Neuilly‐sur‐Seine from January to mid‐February 2021. In the U.S., it was held in Teaneck, New Jersey, from November 2021 to February 2022. At the beginning of the session, each woman cleaned her face and lips using makeup remover, toner, and thermal water to eliminate any residual makeup, to ensure the measurements were taken on bare lips. The color was recorded on the lower lip within defined areas of an 8.2mm² diameter.

2.3.2. Analysis

The hyperspectral images obtained from the SpectraFace were initially calibrated using white and black reference measurements. Following this, color reconstruction was conducted under the D65 illuminant and 10° observer settings, and then computed. This color reconstruction was converted to RGB for visualization and to Lab for parameter computation. Colorimetric analysis was then performed on the Lab color reconstruction within the defined regions of interest. The region of interest was the darker part of the lower lip because it is considered to reflect the true lip color more accurately, particularly for voluminous lips, where the lightest color often represents the inner section of the vermilion (Figure 1). The SpectraFace allowed the extraction of the standard colorimetric parameters of lightness (L*), chroma (C*) and hue (h), which were expressed in the CIE 1976 standard colorimetric space: L*, a*, b*, C*, h.

FIGURE 1.

Regions of interest defined on inferior lips of SpectraFace images, respectively on homogeneous lip color and on heterogeneous lip color (left image and right image).

2.4. Statistics

We deployed Hierarchical Ascending Classification (HAC), a well‐established method of cluster analysis, to determine clusters among participants based on their colorimetric parameters. HAC constructs a hierarchy of clusters by initially considering each observation as an individual cluster. Subsequently, it amalgamates pairs of clusters until all clusters merge into one comprehensive cluster incorporating all of the observations. This process culminates in a dendrogram, a tree‐based representation of observations. Utilizing Euclidean distance and Ward's method, HAC offers a visually rich summary of the clustering procedure, facilitating data structure interpretation. This method, well‐documented in several statistical and methodological references, such as, ¹² was applied with two distinct strategies in our study.

In the first strategy, clusters were identified within the entire population. In the second strategy, clusters were recognized within each ethnic group, followed by cluster identification based on the central observation of these clusters. Subsequently, the nearest clusters among ethnicities were amalgamated. We utilized JMP, Version 15. SAS Institute Inc., Cary, NC, USA, 1989–2023 for our analysis.

3. RESULTS

Figure 2a,b offers a comparative analysis of data across the three ethnicities on two separate planes, Cab*,L* and hab‐L*, on bare inferior lips. The measurements taken from the 410 volunteers (100 Caucasian French, 107 Caucasian American, 103 African American, and 100 Hispanic American women) allowed us to define an expansive continuous lip color space. The maximum difference in lightness was 29.8 (minimum L* = 24.6, maximum L* = 54.4, mean L* = 41.9, standard deviation L* = 6.6), and the hue ranged significantly from red to yellow with a maximum gap of 27.4 (minimum h = 20.8, maximum h = 48.2, mean h = 34.0, standard deviation h = 4.7). With regards to chroma, a considerable range was observed, particularly for African American women, with a maximum gap of 30.2 (minimum C* = 4.9, maximum C* = 35.1, mean C* = 23.6, standard deviation C* = 6.0). Scatter plots indicate that darker lips tend to be less saturated and more yellow, while lighter lips are more saturated and redder.

FIGURE 2.

Inferior lip color space of the 410 volunteers, according to the three ethnicities (A, African American, C, Caucasian (French and American) and H, Hispanic American) in Cab*‐L* (A) and hab‐L* planes (B).

Although there were overlaps, particularly between Hispanic American and Caucasian (French and American) women concerning chroma, lightness, and hue, distinct ethnic differences were noticeable. African American women exhibited the darkest lips, which were also less saturated and more yellow (mean L* = 33.2, mean C* = 16.2, mean h = 36.7), while Caucasian women had lighter, more saturated, and redder lips (mean L* = 46.0, mean C* = 26.7, mean h = 31.9). The lip attributes of Hispanic American women were intermediate for lightness, chroma, and hue (mean L* = 42.3, mean C* = 25.0, mean h = 35.6).

Among the total variability, African American women represented 45%, Caucasian women 23%, and Hispanic American women 32%, indicating a higher diversity amongst African American women, especially concerning chroma and lightness. The broad range of lip color within each ethnic group leads to overlaps between groups. HAC, based on L*, C*, and h data, aimed to differentiate clusters among the 410 women according to their inferior lip color, irrespective of their ethnicities. This analysis resulted in the establishment of nine distinct groups (Table 1).

TABLE 1.

Summary of colorimetric means and standard deviation by HAC‐group from 1‐factor Anova (Group) and Newman‐Keuls post hoc tests.

	Ethnic groups concerned
Group	Number of African American women	Number of Caucasian women	Number of Hispanic American women	Mean L*	Mean C*	Mean h	Color of the lower lip of the group
Group 1	47		2	28.9 d	10.8 f	36.3 c
Group 2	34		12	34.9 c	19.2 e	36.5 c
Group 3	11	1	16	42.3 b	23.4 d	42.6 a
Group 4	3	28	17	47.1 a	25.6 c	38.5 b
Group 5	3	21	19	43.7 b	27.2 b	34.9 d
Group 6	5	53	16	42.3 b	24.9 cd	30.1 e
Group 7		29	6	43.3 b	29.7 a	26.1 f
Group 8		37	5	47.3 a	29.2 a	30.5 e
Group 9		38	7	48.7 a	24.4 cd	33.8 d
F				310.8	282.3	238.8
Pr > F				< 0.0001	< 0.0001	< 0.0001

Note: Two different letters indicate a significant difference for the same variable at the 5% level.

Abbreviations: A, African American; C, Caucasian; H, Hispanic American.

The nine groups presented in Table 1 demonstrated equal representation of subjects, with each group encompassing between 7% and 18% of the total participant count. African American and Caucasian women appeared in six of the nine clusters. Specifically, African American women were absent from clusters 7, 8, and 9, primarily populating clusters 1 and 2, whereas Caucasian women were absent from clusters 1, 2, and 3 and dominated clusters 6, 7, 8, and 9. In contrast, Hispanic American women were present across all clusters, predominantly featuring in clusters 3, 4, and 5. This cluster analysis implies that the vast diversity observed, particularly among African American women, did not fully encapsulate the representation of diversity among Caucasian women. The observed imbalance might partially be attributed to the fact that our sample contained twice as many Caucasian women as African American women, thus necessitating another statistical approach to address this discrepancy (strategy 2).

The subsequent HAC analysis aimed to distinguish unique lip color groups within each of the three ethnicities to characterize the lip color attributes of each group. Following the outline of strategy 2, a second HAC analysis was conducted to gain a more nuanced understanding of lip color overlaps between ethnicities. This analysis amalgamated groups exhibiting similar lip color, culminating in the formation of 11 distinct groups (Table 2). Groups 1, 2, 3, and 5 were solely constituted of African American women, groups 8 and 10 exclusively of Caucasian women, and group 11 solely of Hispanic American women. The remaining four groups were ethnically mixed, among which groups 6, 7, and 9 were combinations of Caucasian and Hispanic American women, and group 4 was an amalgamation of African American and Hispanic American women. The analysis did not yield any mixed group comprising African American and Caucasian women, emphasizing the point that Hispanic American women constitute the ethnicity with the most overlaps.

TABLE 2.

Summary of colorimetric means and standard deviation by HAC‐group from 1‐factor Anova (Group) and Newman‐Keuls post hoc tests.

	Ethnic groups concerned
Group	Number of African American women	Number of Caucasian women	Number of Hispanic American women	Mean L*	Mean C*	Mean h	Color of the lower lip of the group
Group 1 (A1)	37			28.5 g	10.1 g	37.1 bc
Group 2 (A2)	12			30.3 f	13.2 f	33.6 d
Group 3 (A3)	17			33.5 e	17.0 e	38.0 b
Group 4 (A4 + H2)	17		25	42.2 c	23.5 c	40.7 a
Group 5 (A5)	20			36.7 d	22.5 c	33.3 d
Group 6 (C1 + H4)		35	17	43.0 c	28.8 a	27.4 f
Group 7 (C2 + H3)		41	17	42.5 c	26.3 b	32.3 d
Group 8 (C3)		37		47.9 a	29.6 a	30.7 e
Group 9 (C4 + H1)		62	25	48.0 a	26.0 b	36.4 bc
Group 10 (C5)		32		45.6 b	22.3 c	30.4 e
Group 11 (H5)			16	35.7 d	19.9 d	35.7 c
F				294.7	263.4	97.1
Pr > F				< 0.0001	< 0.0001	< 0.0001

Note: Two different letters indicate a significant difference for the same variable at the 5% level.

Abbreviations: A, African American; C, Caucasian; H, Hispanic American.

Upon comprehensive comparison of the two strategies, it was discerned that the second strategy, which resulted in the creation of 11 groups, provided a more accurate depiction of the diversity within the three ethnicities, particularly in the case of African American women (Figure 3a–d). This approach seemed to encapsulate the broad range of lip color variation more comprehensively, ensuring an enhanced representation of ethnic diversity in the study.

FIGURE 3.

Inferior lip color space of the 410 volunteers, according to the two clustering process (A: African American, C, Caucasian, and H, Hispanic American) in Cab*‐L* (a and b) and hab‐L* planes (c and d) with 95% confidence ellipses.

4. DISCUSSION AND CONCLUSION

With an interest in understanding and adapting cosmetic responses to lip color, our work aimed to illuminate the diversity and distinct characteristics of lip color across three ethnicities.

Our study underscored the continuous nature of lip color, marked by a broad spectrum of diversity both within and across ethnicities. Our findings revealed that darker lips tend to exhibit reduced saturation and a yellowish tint, while lighter lips are characteristically redder and more saturated. African American women predominantly exhibited darker, less saturated, and more yellowish lips, whereas the lips of Caucasian women were lighter, more saturated and redder. Hispanic American women represented intermediate values for lightness, hue, and saturation. While chroma and lightness were distinguishing features across ethnicities, hue displayed more overlap, partly due to the immense intra‐ethnic diversity. African American women, in particular, displayed significant diversity, especially in terms of chroma and lightness, a finding consistent with skin color variations reported in previous studies for this specific ethnic group. ¹³

Contrary to skin color, lip color diversity has not been extensively studied. Yet, as highlighted in the work of Baras and Caisey, ⁴ lip color and skin color spaces are vast and complementary, with variations across ethnicities that parallel our own findings. Regarding lip color differences between ethnicities, their conclusions were similar to our findings, with the African American and Indian women in their study having the darkest and less saturated lips, and the Asian and Caucasian women showing the lightest and most saturated lip color. For hue, the Caucasian women had redder lips than the Asian women, and the Indian women had the most yellowish lips. The second part of their study focused on the contrast between skin and lip color without delving into intra‐ethnic diversity. Our study, however, strived to consider the high intra‐ethnic diversity and inter‐ethnic overlaps while measuring lip color. Our clustering analysis yielded 11 groups, providing a better representation of the diversity, especially among African American women. Among these 11 groups, four comprised mixed ethnic groups. These 11 groups are differentiated on all variables (lightness, hue and chroma) to yield a gradient of the lip color. Although our two‐step statistical analysis procedure initially categorized women by their ethnic group (which could be considered contradictory to our goal to go beyond ethnicities), it ensured equal weighting to all ethnicities despite the discrepancy in panel composition. Importantly, our results demonstrate that it is impracticable to link one or more lip colors exclusively to a particular ethnicity. Instead, a single lip color could span multiple ethnic groups, mirroring findings reported for skin color variations. ¹³

Our study is limited by the absence of data from other ethnic groups, particularly Asians. Considering the significance attributed to lip color in traditional Chinese medicine, ¹⁴ and its integration into Chinese quality of life questionnaires, ¹⁵ further studies should incorporate these groups to better understand their specificities and diversity.

Additionally, studies suggest that lip color tends to darken and become more yellow with age, ³ prompting investigation of age‐related variations in lip color across ethnicities. Complementary characteristics of lips, such as hydration, volume, relief, and elasticity, could provide further insights into the intricate nature of lip appearance. As the lips are a significant feature of the face, playing an important role in diverse perceived characteristics such as beauty or age, ¹⁶ , ¹⁷ , ¹⁸ taking into consideration skin and lip color contrasts, with color associated modifications through make‐up, could be another way to explore lip color.

Moreover, it would be beneficial to study the impact of certain cosmetic products or aesthetic procedures that can temporarily alter lip color. ¹⁹ Notably, we propose that the 11 groups of lip colors identified in this study could correspond to 11 nude shades of lipsticks. However, the consumer's definition of a “nude” lipstick shade may not exactly match their lip color and could be slightly different. Further exploration of these nuances would deepen our understanding of lipstick usage and selection across different lip colors.

In conclusion, our comprehensive study of lip color and innovative clusterization for lip color application has identified 11 distinct lip colors, providing an exhaustive representation of lip color diversity across three ethnic groups. These 11 colors could potentially serve as target shades for the development of a more diverse and inclusive range of lip cosmetics, such as nude lipsticks.

CONFLICT OF INTEREST STATEMENT

H.V., Z.C., E.L., and G.G. are full‐time employees of LVMH Recherche, a company dedicated to cosmetic products. V.C. and M.L.F. are full‐time employees of Parfums Christian Dior, a company dedicated to Dior cosmetic products. L.C. is an employee of Boostinnov, a company dedicated to cosmetic product innovation. D.B. is a member of the Université Paris‐Saclay, INRAE, AgroParisTech, UMR SayFood. This work was supported by Parfums Christian Dior and LVMH Recherche.

Vergnaud H, Charton Z, Blumenthal D, et al. Lip color diversity: An intricate study. Skin Res Technol. 2024;30:e13583. 10.1111/srt.13583

DATA AVAILABILITY STATEMENT

The author has provided the required Data Availability Statement, and if applicable, included functional and accurate links to said data therein.