A novel all‐in‐one dynamic wrinkle evaluation system: Development, validation and application in cosmetic industries

Abstract

Objective

This study aimed to develop and validate a novel, integrated dynamic wrinkle assessment system based on standardized, real‐time and multi‐angle facial images and video recordings, and to explore its potential application in evaluating the anti‐wrinkle efficacy of skincare products.

Methods

A standardized protocol was developed for a novel system integrating a high‐precision positioning apparatus with synchronized, multi‐angle high‐speed cameras to capture facial expressions from a neutral to a maximal state. To ensure longitudinal consistency, a novel algorithm was used to select analysis frames based on angular congruence. The system's reproducibility and accuracy were validated by expert dermatologists using the intraclass correlation coefficient (ICC). The system's feasibility and effectiveness were assessed in an 8‐week anti‐wrinkle cream efficacy trial.

Results

The integrated dynamic wrinkle evaluation system demonstrated high inter‐rater reliability (ICC >0.70) across 35 Chinese women, with ICC values exceeding 0.900 for key facial areas. In an anti‐aging efficacy trial (n = 42), the system successfully detected and quantified significant improvements (p < 0.05) in both static and dynamic wrinkles following product use.

Conclusion

This study successfully developed and validated an all‐in‐one dynamic wrinkle photo grading evaluation system for personalized assessment of treatment outcomes following topical application of cosmetic products. The high accuracy of the clinician evaluations offers a new approach for assessing the effects of anti‐aging skincare and cosmetic products, surpassing traditional static assessment methods.

Photo grading of dynamic lines in the dynamic line repeatability validation study.

INTRODUCTION

Humans possess a rich repertoire of facial expressions that reflect emotional states in daily life. The generation of facial expressions involves the contraction of underlying musculature, which exerts tension on the overlying skin and gives rise to transient wrinkles in regions predisposed to folding. These temporary, motion‐induced creases—commonly referred to as expression lines—such as glabellar frown lines, crow's feet and horizontal forehead lines, typically diminish when the expression resolves [1]. During facial expression generation, dynamic wrinkles transiently appear in response to muscular activity, distinguishing them from static wrinkles that persist at rest and predominantly manifest in aging populations. The progression from dynamic to permanent wrinkles occurs through repetitive muscular contraction combined with dermal structural deterioration. For example, while dynamic wrinkles over the glabella, forehead and lateral canthal areas are caused by muscle hyperactivity in the region, these dynamic wrinkles become static wrinkles after years of repeated contraction of the upper facial muscle groups, exacerbated by the loss of collagen and elastin in the dermis [2]. In childhood, the skin is free of wrinkles. However, younger adults (i.e. individuals aged 18–35), who present early signs of aging or ‘premature aging’, may exhibit dynamic lines in response to repeated expressions or tactile gestures. As the skin ages, its resilience decreases due to diminished collagen and elastin production [3, 4, 5], which lead to epidermal thinning and reduction of skin elasticity, whereas the loss of subcutaneous fat, gravitational descent and ongoing reduced capacity to repair micro‐damage from continuous contractions contribute to the formation of both dynamic and permanent wrinkles [6]. This leads to the transformation of dynamic wrinkles into permanent static lines [7]. Dynamic lines have also been proposed as early indicators of future persistent wrinkle development [8].

The desire to maintain a youthful appearance has driven the growth of both cosmetic dermatological procedures and the skincare industry. Aesthetic interventions and topical regimens aim to slow the aging process and rejuvenate the aging face [9]. Botulinum toxin is one of the most widely used injectables. Injections of botulinum toxin reduce dynamic wrinkles by relaxing the facial muscles involved in expression [10, 11], whereas fillers, such as hyaluronic acid, address both dynamic lines and volume loss by augmenting different tissue compartments [12, 13], thereby delaying the progression to permanent wrinkles. Topical skincare products containing active ingredients—most notably anti‐wrinkle creams—are also favoured by consumers. Given the variability in facial aging among individuals, the accurate evaluation of cosmetic interventions requires personalized, comprehensive facial assessments that include both static and dynamic wrinkle analysis [14].

Several internationally recognized rating scales and atlases have been developed to evaluate static wrinkles and to support the assessment of facial aging and cosmetic product efficacy [13, 15]. These include Skin Aging Atlas [15], the Glogau Scale [16], Fitzpatrick Wrinkle Scale [17] and the Crow's Feet Grading Scale [18], which along with the Global Subjective Skin Aging Assessment (GS2A2) [19], provide standardized methods for quantifying wrinkle severity and track the effects of cosmetic products or interventions. In addition, digital imaging techniques have been employed to quantify changes in facial landmarks—such as brow movement—before and after treatment, thereby reflecting improvements in expression lines [20]. However, the evaluation of dynamic wrinkles remains relatively underdeveloped. Some studies have utilized face painting in combination with three‐dimensional skin vector displacement analyses to indirectly assess changes in facial musculature and dynamic wrinkles [21]. Machine learning (ML) has also been introduced in determining skin aging using image‐based and non‐image‐based techniques. A recent systematic review compared the accuracy and limitations of ML‐based techniques [22]. Image‐based techniques include the analysis of full face, eyes and dynamic facial expressions, providing insight for drug development, therapeutic targets and skin aging projections [22]. These approaches highlight the importance of integrating dynamic facial expression analysis tools for dermatologists. Such tools promise to provide more accurate and personalized measurements of skin aging for patients.

Although contour‐fitting methods for facial expression modelling are widely used in facial recognition and facial expression recognition (FER) technologies [23], a standardized and individualized objective approach for evaluating dynamic wrinkles is lacking in the cosmetic field. This highlights the need for the development of an integrated, objective tool capable of capturing real‐time changes in dynamic facial lines, along with corresponding clinical grading. The aim of this study was to develop and validate an integrated dynamic wrinkle assessment system capable of analysing facial expressions and dynamic lines based on standardized, real‐time and multi‐angle facial images and video recordings. Through reproducibility testing by board‐certified dermatologists and application in an anti‐aging efficacy study involving an Eye Cream for the Face, the system's feasibility and effectiveness in both clinical settings and cosmetic efficacy evaluations were preliminarily explored.

METHODS

Development and validation of an all‐in‐one dynamic wrinkles evaluation system

Study design and participants

This study comprised two components: the development of an integrated dynamic wrinkle assessment system and the evaluation of its reproducibility and accuracy.

The development phase was designed to create a standardized protocol for dynamic wrinkle assessment, including optimized imaging settings, software configuration and various standardized facial expression instructions for users. During this phase, a 40‐year‐old female volunteer of Han Chinese ethnicity was enrolled, who had extensive experience with cosmetic products and no dermatologic or systemic conditions that could potentially interfere with the acquisition of dynamic facial imagery.

Reproducibility validation was conducted in Shanghai between 25 October and 3 November 2023. This phase enrolled Chinese females aged 20–45 years with all skin types, self‐reporting lack of elasticity and firmness, as well as rough skin. Eligible participants typically displayed expressive facial behaviour—frequent smiling, eyebrow raising and frowning. Additional inclusion criteria required participants to present with specific wrinkle grades, as assessed by dermatologists using the Skin Aging Atlas [15]. These include Smile expression—Underneath eye wrinkles: grades 3 to 9; crow's feet: grades 3 to 6; nasolabial folds: grades >1 to 7; eyebrow‐raising expression—forehead wrinkles: grades 3 to 8; frown expression—glabellar lines: grades 2 to 6. Clinical assessments were performed by three board‐certified dermatologists with 17, 34 and 36 years of professional experience, respectively. Detailed inclusion and exclusion criteria are provided in the Supporting Information. Written informed consent was obtained from all participants.

All‐in‐one dynamic wrinkles evaluation system components

The integrated wrinkle evaluation system consisted of three components: a positioning apparatus (custom‐made ear clips and metallic frame), an image acquisition system (hardware: high‐speed cameras; software: LUMIX) and an image analysis module (Figure 1). The system enabled real‐time, multidirectional capture of facial dynamic wrinkles in both photo and video formats during facial expressions. The positioning device comprised ear clips configured to hold the head and ears of the user, and a height ruler configured to define the height of the head of the user. The image acquisition system incorporated three high‐speed cameras (Panasonic, GH5, Japan) and a primary computational unit. Cameras were placed in front of and on both sides (left and right) of the participant. Image capture was facilitated by proprietary LUMIX software, which recorded synchronized high‐resolution images from all three perspectives. The entire acquisition was conducted under a standardized light environment with fixed camera parameters to ensure uniformity: ISO was set to 400, focal length to 16 mm, aperture to f/5.6, shutter speed to 1/80 s and light temperature to 6500 K. The image analysis module consisted of a computing device with memory storage and a processor. The memory stored acquired images, while the processor—equipped with a secondary computational circuit—processed the images for wrinkle analysis. This module communicated wirelessly with the image acquisition system and provided real‐time feedback to guide user repositioning via the positioning apparatus, thereby improving image consistency.

FIGURE 1.

Configuration of dynamic wrinkle evaluation device.

All‐in‐one dynamic wrinkles evaluation system operating procedure

The overall operating procedure for the dynamic wrinkle assessment is outlined in Figure 2. Prior to image capture, the volunteer's head and ears were positioned in a stabilized position, with head height standardized using the positioning apparatus. For follow‐up visits, the volunteer was repositioned to replicate the exact orientation of previous image captures, allowing real‐time alignment and fold‐over analysis across time points.

FIGURE 2.

Flow diagram for dynamic wrinkles evaluation on keratin material.

To obtain concurrent images, participants were first required to watch the user instruction video and undergo training for the entire process. Next, trained technicians captured a continuous series of images of facial expressions using high‐speed cameras while participants performed three facial expressions—smiling, raising the eyebrows and frowning—progressing from a neutral to a maximal state. Neutral expression was defined as a resting facial state without any voluntary muscle activation, whereas maximal expression referred to the greatest degree of facial contraction elicited under technician instruction. During each expression task, images were simultaneously acquired from the front, left and right perspectives using high‐speed cameras, with each recording session lasting approximately 3 s. Each camera generated approximately 100 to 120 images per session. The captured images were stored and processed to evaluate dynamic wrinkles using a strictly predefined criteria. Specifically, a standardized 10‐image series at T0 was manually curated along the expression continuum from neutral state to peak expression intensity (Figure 3). In subsequent time points during facial expression quantification, automated frame selection employed continuous measurement of the angle formed by the nasal tip and bilateral oral commissures. The computational algorithm enforced strict angular congruence across selected frames relative to T0 values.

FIGURE 3.

The standardized 10‐image series from neutral to peak expression at T0 and T8w.

Validation of the all‐in‐one dynamic wrinkles evaluation system

High‐speed camera imaging acquisition was performed at 4 h (T4h, 1 p.m.) and 8 h (T8h, 5 p.m.) after baseline (T0, 9 a.m.) to capture the three facial expressions of the participants (smiling, raising of eyebrows and frowning) from a neutral to maximal expression state. At T0, T4h and T8h, dynamic wrinkle grading was conducted by three dermatologists independently based on the images.

The outcomes included photo grading of dynamic wrinkles at each time point and computation of the intraclass correlation coefficient (ICC) among dermatologists.

Dynamic wrinkles grades were assessed using Skin Aging Atlas [15], focusing on the features established below for each expression–smiling: underneath eye wrinkles, crow's feet wrinkles and nasolabial folds; eyebrow raising: forehead wrinkles; frowning: glabellar wrinkles. Three external experts independently graded the photographs captured. The mean clinical score based on 10 PCS images obtained throughout the expression process was calculated.

Anti‐aging clinical study of an anti‐wrinkle cream in healthy participants

The system's practical utility and effectiveness were assessed in a single‐arm clinical study conducted from 17 November 2021 to 30 January 2022, in Shanghai, spanning the late autumn and winter seasons. To minimize the influence of external environmental fluctuations on data acquisition, all instrumental and photographic assessments at each visit were performed after a 30‐min acclimatization period in a strictly controlled room with constant temperature (21°C ± 1°C) and humidity (45% ± 5%). Efficacy assessments, including dynamic wrinkle analysis with the novel system, were conducted at baseline, 15 min (Timm) and at weeks 4 and 8. The full ingredient list of the investigational product, according to the International Nomenclature of Cosmetic Ingredients (INCI), is provided in Supporting Information: Section 1. The detailed design, including inclusion and exclusion criteria and a detailed list of all assessments, is provided in Supporting Information: Sections 4 and 5.

The study was conducted in accordance with the Declaration of Helsinki and the guidelines for Good Clinical Practice of the International Council for Harmonization. Ethics approval was obtained from the Shanghai Ethics Committee for Clinical Research (SECCR/2021‐146‐02), and written informed consent was provided by all participants.

Statistical analysis

The analysis set for the reproducibility validation study included all participants who completed the required study visits. The ICC for dynamic wrinkle assessments by different experts was calculated [24]. If ICC was between 0.6 and 0.8, it indicated a strong correlation, whereas an ICC between 0.8 and 1 indicated a very strong correlation.

The efficacy analysis set (EAS) for the Anti‐Aging Clinical Study of anti‐wrinkle cream comprised all participants who used the study product at least once and completed the required visits. The Safety Analysis Set (SS) for the Anti‐Aging Clinical Study of anti‐wrinkle cream included all participants who used the study product at least once and completed the 4‐week safety questionnaire. Changes from the baseline of the efficacy endpoints were analysed using one‐way ANOVA for normally distributed variables or the Wilcoxon signed‐rank test otherwise.

All statistical tests were two‐sided, and a p‐value of <0.05 was considered statistically significant. Analyses were conducted using SPSS software, v.20 (IBM Corp., Armonk, NY, USA).

RESULTS

Validation of all‐in‐one dynamic wrinkles photo grading evaluation system

Participant characteristics

Thirty‐five female Asian participants were initially enrolled in this study. Thirty participants completed the study and were included in the final analysis set. The mean age of participants was 38.33 years (SD 4.21). Among them, 23 (76.7%) reported having sensitive skin, whereas 7 (23.3%) had non‐sensitive skin. The skin types were distributed as follows: dry skin, 8 (26.7%); normal skin, 7 (23.3%); oily skin, 2 (6.7%); and combination skin, 13 (43.3%) (Table 1).

TABLE 1.

Characteristics of participants in the dynamic line repeatability validation study.

Variables	Participants (N = 30)
Age (years), mean ± SD	38.33 ± 4.21
Ethnicity, n (%)	30 (100)
Female, n (%)	30 (100)
Skin type, n (%)
Dry	8 (26.7)
Normal	7 (23.3)
Oily	2 (6.7)
Combination	14 (43.3)
Sensitive skin, n (%)
Yes	23 (76.7)
No	7 (23.3)

Photo grading dynamic lines

When comparing the results to the baseline, no significant changes were observed in crow's feet wrinkles, under‐eye wrinkles or nasolabial folds at T4h and T8h when assessed with a smiling expression. Similarly, there were no notable changes in forehead wrinkles during a raised eyebrow expression or glabellar wrinkles during a frowning expression at T4h and T8h (Figure 4).

FIGURE 4.

Photo grading of dynamic lines in the dynamic line repeatability validation study.

Consistency of expert photo grading

The ICC for the three expert clinical assessments for the dynamic wrinkles' images at different follow‐up time points was evaluated. For all expression modes, the clinical photo grading scores of the three experts relative to the dynamic wrinkles had an ICC >0.700. For the smile expression, the ICC of the clinical experts following photo grading for crow's feet wrinkles at the different follow‐up time points was >0.600, whereas the ICC of expert graders for the underneath eye wrinkles and nasolabial folds was >0.900. For the raised eyebrow expression, the ICC for the expert clinical grading scores relative to the forehead wrinkles at the different follow‐up time points was >0.900. Similarly for the frown expression, the expert clinical grading scores for glabellar wrinkles at the follow‐up time points were >0.800 (Table 2).

TABLE 2.

Consistency.

Attributes	Expression parameter	Expert 1			Expert 2			Expert 3			All
		T0 vs. T4h	T0 vs. T8h	T4h vs. T8h	T0 vs. T4h	T0 vs. T8h	T4h vs. T8h	T0 vs. T4h	T4h vs. T8	T4h vs. T8h	T0 vs. T4h	T4h vs. T8	T4h vs. T8h
Smile expression
Crow's feet wrinkles	Max	0.858	0.646	0.756	0.882	0.717	0.757	0.857	0.738	0.857	0.886	0.717	0.804
	Mean	0.906	0.842	0.889	0.939	0.859	0.893	0.857	0.790	0.902	0.909	0.844	0.898
Underneath eye wrinkles	Max	0.959	0.979	0.979	0.970	0.974	0.981	0.988	0.982	0.993	0.978	0.981	0.992
	Mean	0.947	0.963	0.957	0.958	0.961	0.961	0.963	0.970	0.975	0.958	0.969	0.967
Nasolabial folds	Max	0.989	0.981	0.981	0.960	0.956	0.969	0.974	0.972	0.993	0.987	0.984	0.989
	Mean	0.938	0.943	0.942	0.923	0.924	0.945	0.947	0.967	0.966	0.940	0.954	0.961
Raised eyebrow expression
Forehead wrinkles	Max	0.979	0.984	0.988	0.993	0.978	0.988	0.963	0.948	0.985	0.986	0.979	0.993
	Mean	0.942	0.957	0.935	0.943	0.956	0.937	0.949	0.955	0.945	0.943	0.956	0.939
Frown expression
Glabellar wrinkles	Max	0.965	0.956	0.982	0.983	0.958	0.981	0.982	0.877	0.891	0.983	0.959	0.977
	Mean	0.960	0.963	0.952	0.954	0.954	0.959	0.951	0.946	0.941	0.958	0.958	0.953

Anti‐aging (anti‐wrinkle efficacy) study

Participant characteristics

Forty‐two participants completed all required visits and were included in the EAS and SS. The mean age of the participants was 42.21 years (SD = 3.60). A full description of the participants is provided in Supporting Information: Section 6 (Table S1).

Efficacy

Following the use of the anti‐wrinkle cream, statistically significant reductions from baseline (p < 0.05) were observed in static wrinkle scores at all follow‐up visits. The intensity of dynamic wrinkles during expression was also significantly reduced from baseline at all follow‐up visits (p < 0.05). Instrumental measurements of skin elasticity and 3D wrinkle parameters also showed statistically significant improvements from baseline (p < 0.05).

Detailed results are provided in the Supporting Information: Section 6. This includes clinical grading for static lines (Table S2), skin elasticity and wrinkle parameters (Table S3), self‐assessment questionnaire data (Table S4) and photo grading for dynamic lines (Figure S1).

Safety

No AEs or serious AEs were reported throughout the study period.

DISCUSSION

This study developed and validated an individualized dynamic wrinkle assessment system applicable for the evaluation of changes in dynamic wrinkle changes for the same individual at different anti‐wrinkle treatment time points or for comparisons before and after the use of other cosmetics, using various facial expression patterns. Clinical assessments performed by experienced dermatologists using this photo grading system demonstrated excellent repeatability. The developed photo grading system was then evaluated in a clinical anti‐aging efficacy study in which an anti‐wrinkle cream was applied to the face. The results showed that the treatment not only significantly improved skin elasticity and other parameters but also enhanced both clinical grading of dynamic and static wrinkles. These findings suggested that the anti‐wrinkle cream treatment may improve static wrinkle progression by influencing the development of dynamic wrinkles.

The all‐in‐one dynamic wrinkle photo grading evaluation system developed in this study provides a new dimension for assessing the effects of anti‐aging skincare and/or cosmetic products, surpassing traditional static assessment methods. Previous anti‐aging/wrinkle assessment methods have mainly focused on static evaluations of biomechanical parameters of the skin, such as wrinkle depth (Ry) obtained by optical profilometry, or skin elasticity measured by the suction method whereby resistance of the skin to the negative pressure (firmness) and its ability to return into its original position (elasticity). However, these methods are often biased and influenced by subjective measurement [25]. A more advanced approach to assessing the improvement of periorbital wrinkles involved the three‐dimensional Fast Optical in vivo Topometry of Human Skin (FOITS) device. The FOITS is a non‐invasive projection technique that collects 3D information from the skin surface to assess average roughness (Ra) and average roughness depth (Rz). However, FOITS uses surface roughness parameters defined in the metalworking industry (such as Ra, Rz and Rmax), which may not fully apply to describing skin characteristics. More importantly, FOITS focuses on static image acquisition and may not be sensitive enough for progressive changes that require long‐term observation. The Visual Illusion‐based Image Feature Enhancement System (VIS) allowed the assessment of wrinkles in real time. The VIS, which relies solely on standard photographic equipment and image processing systems, is susceptible to environmental influences, particularly during dynamic wrinkle assessments [26]. However, variations in facial expression and lighting can introduce errors, which may affect the accuracy of the evaluation. The developed all‐in‐one dynamic wrinkle evaluation system in this study captures the dynamic characteristics of facial wrinkles formation in response to facial expressions. Independent evaluations by experienced dermatologists using this dynamic wrinkle assessment system demonstrated excellent repeatability and consistency. These results validate the effectiveness of this system as a reliable tool for clinical research and underscore its practical application value.

The anti‐aging efficacy study of anti‐wrinkle cream further elucidated the potential link between dynamic and static wrinkles. Specifically, changes in dynamic wrinkles may be an important factor in the development of static wrinkles. Thus, by inhibiting the formation of dynamic wrinkles, it may be possible to effectively prevent or alleviate the occurrence and progression of static wrinkles. Mechanistically, wrinkles are caused by photodamage, squinting, constant movement of the eyes, aging and chronic rubbing due to seasonal allergies. Fujimura and Hotta [7] proposed that the degree of dynamic wrinkles may reflect that of static wrinkles and static wrinkles derive from the kinetic origins of dynamic wrinkles. Over time, the skin loses its elasticity, and wrinkles appear owing to the reduced levels of collagen and elastin fibres. In addition to the natural wrinkles caused by aging (static lines), expression wrinkles form owing to the contraction of facial muscles (dynamic lines).

The primary ingredients of anti‐wrinkle cream in this study include hydroxypropyl tetrahydropyrantriol (Boscaline), dipeptide diaminobutyroyl benzylamide diacetate (Syn‐Ake), fermented complex (Yeast/Bacillus Vulgatus/Black Tea Fermentation Products) and sodium hyaluronate. Previous studies on these key ingredients support their broader anti‐aging efficacy [27]. Hyaluronic acid has been shown in clinical trials to increase skin moisture and reduce wrinkle depth [28]. Peptides and antioxidants have demonstrated synergistic anti‐aging effects in both facial and periorbital application studies. In particular, these components promote collagen production, reduce inflammatory responses and inhibit non‐enzymatic glycation and the formation of AGEs, contributing to the improvement of skin elasticity and expression lines. Anti‐aging peptide compounds have been shown to upregulate collagen production (in vitro) and have shown clinical improvements when applied topically to photodamaged skin [27, 29, 30]. The clinical study evaluating anti‐wrinkle cream provides evidence supporting the relationship between dynamic wrinkles and static wrinkles, and further confirms the value of the all‐in‐one dynamic wrinkle evaluation system in both clinical research and skincare product development. Future research should focus on the role of dynamic wrinkles in the skin aging process and explore more targeted methods and technologies for managing dynamic wrinkles.

This study has several limitations. The primary limitation stems from the single‐arm, open‐label design of the clinical application study. Without a concurrent vehicle control or a split‐face design, it is not possible to definitively attribute the observed improvements solely to the active ingredients, as contributions from the product chassis or placebo effects cannot be excluded. Furthermore, as the study was conducted over a 2‐month period, potential confounding from seasonal variations in skin physiology could not be controlled for. While this design was sufficient for demonstrating the system's feasibility, future trials aimed at definitive claim substantiation should employ a randomized, controlled methodology. The generalizability of the findings is also limited to the study population of Chinese females, and further validation in diverse ethnic groups is warranted. Finally, while performed by experts, the clinical grading remains a subjective endpoint; future work could integrate automated image analysis to enhance objectivity.

CONCLUSION

This study developed and validated an all‐in‐one dynamic wrinkle photo grading evaluation system for personalized assessment of treatment outcomes following typical application of eye wrinkle cream products. The high accuracy of the evaluation by clinicians provides a new dimension for assessing the effects of anti‐aging skincare and/or cosmetic products, surpassing traditional static assessment methods. Further large‐scale studies are needed to expand the generalizability of the approach in different populations. The system shows promise as an important reference tool for skincare product manufacturers and professionals.

AUTHOR CONTRIBUTIONS

Xingzuo Liu: Methodology; investigation; validation; project administration; formal analysis; writing—original draft; writing—reviewing and editing. Juanjuan Chen: Methodology; investigation; project administration; validation; visualization; writing—original draft; writing—reviewing and editing. Andrew Steel: Supervision; writing—reviewing and editing. Frederic Flament: Methodology; conceptualization; writing—reviewing and editing; supervision. Jing Li: Methodology; conceptualization; project administration; investigation; validation; writing—reviewing and editing. Chengda Ye : Methodology; conceptualization; investigation; writing—reviewing and editing. Yidong Tu: Supervision; writing—reviewing and editing.

FUNDING INFORMATION

The study was supported by L'Oréal Research and Innovation.

CONFLICT OF INTEREST STATEMENT

All authors are employees of L'Oréal Research and Innovation, which sponsored the study.

Supporting information

Data S1.

DATA AVAILABILITY STATEMENT

All data on the measured ecosystem variables indicating ecosystem functions that support the findings of this study are included within this paper and its Supporting Information files.