Sign in

Quantifying Visual Aspects of Hair

Contact Author Trefor Evans, Ph.D., TRI-Princeton, Princeton, New Jersey
Fill out my online form.
Woman with voluminous hairstyle

An old adage suggests that “a picture is worth a thousand words.” In an analogy, it is very easy to recognize beautiful hair when we see it; but perhaps not so easy to describe the qualities that make it so. The hair measurement scientist is further tasked with translating this concept into numbers.

Past articles in this column have discussed means of performing this undertaking: We have considered sensorial properties, such as surface friction and associated manageability issues;1 deliberated on various aspects of hair health;2-4 covered measurement of hair’s physical properties;5, 6 and the detrimental effects of certain habits, practices and insults.7, 8 Unquestionably, these are all contributors but at the same time, it is likely that visual perception is probably most important of all. Unfortunately, quantifying the visual properties of hair is an especially taxing proposition.

As scientists, we are able to conceive of highly controlled, precise evaluation methods to probe our technical interpretations of consumer terms. To this end, previous articles in this series have described how the measurement of light reflecting off a hair surface is commonly used as a measure of “shine”;9 or how image analysis approaches are often used to quantify hair “volume.”10 These approaches undoubtedly hold up in terms of scientific rigor but at the same time, relationships back to the consumer terms are tenuous.

Excerpt Only This is a shortened version or summary of the article you requested. To view the complete article, please log in or create an account. Registration is Free!

Many hair care properties possess a means of technical evaluation that equate to expectations relating to consumer language. However, the same is not true for most evaluations of visual properties.

The reason for this likely is twofold. First, as per a reoccurring theme in these articles, consumer language can often be rather nebulous, and may not equate to scientific understanding. But second, from a technical perspective, these simple consumer terms represent extremely complex phenomena.

By means of illustration, even something as seemingly straightforward as measuring hair color is fraught with complexity. Spectrophotometers are commonly used for such quantification but at the same time, color has dependence on the nature of the light source and its angle of incidence. Subtle differences in color will be present between neighboring fibers, as well as down the length of a given fiber. To this end, a permanent dye treatment that produces very uniform coloring of hair is thought unnatural due to an absence of contrast.

This article describes how similar complexities plague many such visual evaluations of hair and therefore produce some of the most contentious properties in our industry. Ongoing research efforts are highlighted that aim to improve our understanding.

The Complexity of the Problem

To again paraphrase the words of the illustrious Lord Kelvin (1824-1907): “When you cannot express it in numbers, your knowledge is of a meager and unsatisfactory nature.” Many of the consumer properties discussed in this series (e.g., conditioning, strength, etc.) possess a means of technical evaluation that equate reasonably well to expectations relating to consumer language—and have therefore achieved a high level of acceptance within our industry. However, the same is not true for most evaluations of visual properties.

This situation does not arise from any lack of effort. For example, many equations have been proposed in the scientific literature in an attempt to generate a number that describes hair shine.9 These all include a ratio involving the proportions of specular and diffusely reflected light that bounces off hair’s surface. Each possesses technical merit but very different magnitude benefits (or indeed detriments) arise when treating reflection data via these different equations. Indeed, the rankings of samples may be altered as a result of the equation selected. Accordingly, and unsatisfactorily, current predilections for a specific equation most often involve the one that yields optimal results for the sample being tested.

A drawback of the most commonly used approach for measuring shine—the so-called polarized light method9—is the artificial alignment of hair. That is, test samples are anchored over a drumlike tress holder during evaluation. Yet, fiber alignment is seemingly a major contributor to this property, with light reflecting more-cleanly from flat, homogeneous surfaces. As such, this important consideration is factored out of the current evaluation.

Advancement in this area necessitates adapting the principles of this approach to freely aligned hair tresses; or better still, real heads. Such efforts are ongoing and this work has been presented at recent hair conferences in both Germany11 and the United States.12

Hair Alignment

Hair alignment is a much-overlooked visual consideration. While not necessarily a property that consumers mention, its influence is nonetheless seen in many forms. Aligned hair is likely to appear shinier; grooming is easier; smoothness may be enhanced; frizz is minimized; and hair seemingly moves more freely. Again, this prompts a desire for quantification. One approach involves a different adaptation of polarization imaging.

The polarization of an incident light source can be altered by the angle at which it reflects off of a surface. Therefore, examining the nature of this alteration yields information on the morphology of the reflecting substrate.13 For visualization purposes, these shifts in polarization can then be assigned a color. Figure 1 shows images of hair tresses that were generated using a commercial instrument that operates by this principle. In this specific example, rather frizzy hair was heat-straightened with a flat iron and then allowed to progressively revert to its natural confirmation via storage at elevated humidity. This produced a spectrum of aligned states, which allowed for comparing visual perception to quantification values yielded by the instrument.


Clearly, an array of hair fibers can be described by a variety of properties. Image analysis methods can be employed to describe the size and shape of the hair via pixel counts from digital images. By this means, technical measures of hair volume can be obtained; although, again, it is a stretch to equate this scientific parameter to the complex consumer term.10 In any such analysis, it can be tricky to define the hair edge since outermost strands might have wispy flyaway properties that do not equate to the tress bulk. Such fibers are undesirable and, in sufficient quantity, can lead to a perception of frizz.

These flyaway fibers can be better visualized by back illumination of the hair where they become haloed against the contrast provided by the bulk, as seen in Figure 2.14 Therefore, a degree of flyaway can be evaluated from ratios of pixel counts for the wisps and bulk.

Hair Motion

In addition to the static properties of hair, its motion appears as another key contributor to the perception of beauty. This topic has been rather sparely studied,15 but is an active area of the present author’s ongoing research. Experiments are being performed on a modified version of the equipment described above for evaluating hair shape and flyaway. In the modified instrument, an oscillating side-to-side motion is induced in hair tresses, while a high-resolution video camera records the outcome. Comparable analyses can then be performed on a frame-by-frame basis.

Figure 3 shows illustrative images acquired during a typical experiment. The shape of the hair and the extent of its motion can be tracked and measured under controlled and specific test conditions.

Experiments that best simulate real-life occurrences are more desirable when attempting to communicate the magnitude of a consumer benefit.

The nature of this motion is highly frequency-dependent and, accordingly, experiments across a range of values are believed most insightful. Figure 4 shows results for the amplitude of hair motion as a function of applied frequency for tresses treated with silicone oil, relative to clean hair. Initially, the application of higher frequency leads to higher amplitudes of motion; although this relationship ultimately peaks and reverses. These curves are dampened by the silicone treatment—that is, their height is lessened, while the maximum amplitude is shifted to higher frequency.

Visually, the silicone-treated hair moves in a notably diminished manner. Therefore, it is theorized that one component of improved motion involves the attainment of higher amplitudes—possibly also under the action of lower frequency.

A natural motion also seemingly necessitates some degree of hair shape change (see Figure 3). During an oscillatory path of motion, the tress and its fibers experience differences in momentum. Accordingly, both the shape and the volume of the hair change during this process—and again, the magnitude of these occurrences is highly frequency-dependent. Figure 5 shows this regular “pulsing” of tress volume and demonstrates how this effect is exacerbated by increased frequency.


As measurement scientists, we must be conscious of whether our methods are providing a convenient means of technical characterization, or are attempting to simulate real-life occurrences. In the grand scheme, both have merit. Characterization experiments tend to have higher precision and can help shed light on important contributing variables. For example, despite the stated caveats, traditional shine experiments allow for recognizing the effect of, for example, fiber surface integrity, hair color, surface coatings, etc. Yet, experiments that best simulate real-life occurrences are more desirable when attempting to communicate the magnitude of a consumer benefit.

Considerable effort has been expended on attempts to quantify visual properties of hair, with all still seemingly residing squarely in the realm of “characterization.” Sound scientific approaches allow for the evaluation of numbers representing hair shine, fiber alignment, spatial volume, frizziness, etc., yet their relationship to consumer perception is frequently debated.

To this end, it is worth pointing out that these are highly complex, multifaceted parameters, and to try and describe them with a single number is perhaps overly ambitious. We have been conscious of this idea in our ongoing efforts on hair motion where we produce a plethora of numbers. Our approach generates values that describe the extent of motion (i.e., the amplitude); the shape of the hair during this process; and the degree of homogeneity/flyaway during motion along an oscillatory pathway. These properties alter rhythmically under this stimulus as the hair experiences momentum differences. Moreover, each of these parameters also possess a sizable dependence on the magnitude of the oscillating frequency. Accordingly, clearer insights are obtained from viewing the data in graphical form (such as those shown in Figures 4 and 5), rather than by quoting any specific value.

It is perhaps tempting to invent equations that put a single number on the various visual properties we have discussed, but these expressions have limited relevance without equating to consumer perception. To this end, perhaps advancement begins with the reverse scenario, by which we first analyze consumer perception and then work backward using weight of element statistical modeling to examine the contribution of measured variables. This approach is being used in the previously referenced work that attempts to better evaluate shine on nonaligned hair.11, 12

Quantitative claims (i.e., 5x stronger, 3x shinier) are extremely popular in the marketing of hair care products. Yet, these communication messages can be questioned if the results and approaches do not equate to real life occurrences and observations. Therefore, and in accord with good scientific practices, it is imperative that we continuously question and upgrade our test methods.

Acknowledgements: The images in Figure 1 were generated by Ernesta Malinauskyte, senior scientist at TRI, using equipment loaned to TRI by Bossa Nova Vision (Culver City, Calif.).

Figure 2 was provided by Bossa Nova Vision, and ongoing hair motion research is being performed in collaboration with Sebastien Breugnot of Bossa Nova Vision.


  1. TA Evans, Evaluating hair conditioning with instrumental combing, Cosm & Toil 126(8) 558-563 (2011)
  2. TA Evans, How damaged is hair? Part 1: Surface damage, Cosm & Toil 132(4) 38-48 (2017)
  3. TA Evans, How damaged is hair? Part 2: Internal damage, Cosm & Toil 132(6) 36-45 (2017)
  4. TA Evans, How damaged is Hair? Part 3: Better defining the problem, Cosm & Toil 132(7) 58-67 (2017)
  5. TA Evans, Measuring hair strength, part 1: Stress-strain curves, Cosm & Toil 128(8) 590-594 (2013)
  6. TA Evans, Measuring hair strength, part 2: Fiber breakage, Cosm & Toil 128(12) 854-859 (2013)
  7. TA Evans, Beating the damaging effects of heat on hair, Cosm & Toil 130(5) 28-33 (2015)
  8. TA Evans, The effects of sun on hair, Cosm & Toil 131(7) 46-52 (2016)
  9. TA Evans, Equating the measurement of hair shine, Cosm & Toil 131(1) 28-34 2016)
  10. TA Evans, Hair volume and body–A technical dissertation, Cosm & Toil 133(6) 48-55 (2018)
  11. M Vedel et al, Evaluation of hair gloss on randomly oriented fibers using polarization imaging, Proc of DWI HairS’17 Conference, Dresden, Germany (Sep 2017)
  12. S Stofel, R George and S Breugnot, Evaluation of hair gloss on randomly oriented fibers using polarization imaging, Proc of 8th Intl Conf on Applied Hair Science, Red Bank, NJ (2018)
  13. N Lechocinski and S Breugnot, Fiber orientation measurement using polarized imaging, J Cosmet Sci 62 85-100 (2011)
  14. S Breugnot et al, Volumizing, flyaway/frizz control and straightening claim substantiation using 3D volume measurement system, NutraCos Cosmetics 7(2) 9-14 (2017)
  15. A Galliano, M Lheur and R Santoprete, Analyzing the movement of a hair swatch using video and image analysis: A promising technique for exploring the dynamic properties of hair, Intl J Cos Sci 37(1) 56-62 (2015)

Related Content



Figure 1. Numerical values for hair alignment

Numerical values for hair alignment from a polarized light approach

Figure 2. ID and quantification of frizzy flyaway fibers

Identification and quantification of frizzy flyaway fibers

Figure 3. Different means for analyzing hair during motion

Different means for analyzing hair during motion

Figure 4. Amplitude versus frequency curves

Amplitude versus frequency curves to describe the nature of hair motion

Figure 5. Change in volume of a hair tress

Figure 5. frequency

Next image >