High speed stereo visualization of hair-skin dynamics

High speed stereo visualization of hair-skin dynamics

Citation for published version (APA):

Sridhar, V., & Technische Universiteit Eindhoven (TUE). Stan Ackermans Instituut. Design and Technology of Instrumentation (DTI) (2010). High speed stereo visualization of hair-skin dynamics. Technische Universiteit Eindhoven.

Document status and date: Published: 01/01/2010

Document Version:

Publisher’s PDF, also known as Version of Record (includes final page, issue and volume numbers)

Please check the document version of this publication:

• A submitted manuscript is the version of the article upon submission and before peer-review. There can be important differences between the submitted version and the official published version of record. People interested in the research are advised to contact the author for the final version of the publication, or visit the DOI to the publisher's website.

• The final author version and the galley proof are versions of the publication after peer review.

• The final published version features the final layout of the paper including the volume, issue and page numbers.

Link to publication

General rights

Copyright and moral rights for the publications made accessible in the public portal are retained by the authors and/or other copyright owners and it is a condition of accessing publications that users recognise and abide by the legal requirements associated with these rights. • Users may download and print one copy of any publication from the public portal for the purpose of private study or research. • You may not further distribute the material or use it for any profit-making activity or commercial gain

• You may freely distribute the URL identifying the publication in the public portal.

If the publication is distributed under the terms of Article 25fa of the Dutch Copyright Act, indicated by the “Taverne” license above, please follow below link for the End User Agreement:

www.tue.nl/taverne

Take down policy

If you believe that this document breaches copyright please contact us at:

openaccess@tue.nl

providing details and we will investigate your claim.

H I G H S P E E D S T E R E O

V I S U A L I Z A T I O N O F H A I R - S K I N

D Y N A M I C S

One year project presented to Eindhoven University of Technology towards the degree of Professional Doctorate in Engineering in

Design and Technology of Instrumentation

Main Report

November 2010 Care and Health Applications Philips Research, Eindhoven,

High Speed Stereo Visualization of Hair-Skin Dynamics / by Vijayalakshmi Sridhar Eindhoven, 2010. – (Design and Technology of Instrumentation)

A catalogue record is available from the Eindhoven University of Technology Library ISBN: 978-90-444-1003-7

/Design and Technology of Instrumentation Author Vijayalakshmi Sridhar Date 30th November 2010

High Speed Stereo

Visualization of

Hair-Skin Dynamics

Main Report

V. Sridhar

Supervisors

Technische Universiteit Eindhoven University of Technology

1

ABSTRACT

Philips has a huge electric shaving market and in order to compete at the fore fronts, the shave needs to be as close as possible without compromising on the comfort of the customer. Hair and skin manipulation (HSM) are vital factors that prepare the skin and hair for a smooth and efficient shave. While several efforts have been laid down in quantifying the ‘closeness’ in shave or to find the exact length of the beard before and after shaving, the most popular and confident approach has been a simple user perception. In order to obtain a more confident number it is required to build an optical setup which can both qualitatively visualize and quantitatively derive 3D data to assess the HSM process.

Thus the work described in this project is towards the design and construction of an opto-mechanical setup that can simulate the shaving process and simultaneously capture high-speed videos of the HSM process happening on the beard. Stereo imaging with two high speed cameras was chosen as the working principle of the setup such that it is suitable to extract 3D data from the hairs. Semi automatic software helps in manually detecting the coordinates of the hairs and through it, its length and its angles with respect to skin. The entire setup is calibrated with a dummy hair-skin phantom and an error analysis is done to determine the accuracy of the set up.

In house manipulation concepts were put together as six module geometries each with varying parameters of interest. These modules were mounted on to the 3D setup and similar to shaving speed strokes were made on three regions of interest – cheek, neck and chin. A total of 40 single hairs were targeted (two strokes each) and experiments were done in unmanipulated state and in manipulated state by the modules to see the exact influence of the modules.

The results show the best geometry of the module has 76% hair manipulation efficiency with an average hair lift of 108µm. While cheek is the best place for manipulation, chin is the toughest place as it is hard to follow contour in that region. It is found that the module does not have uniform contact with the skin and this affects the results as the hairs need to be attacked at their root level. Taking into consideration, only the most efficiently manipulated hairs the average hair lift is found to be 105µm and the average remaining stubble (virtual) would be -36µm.

This project lays the foundation for accurately measuring the 3D data of a hair while manipulating hair and skin with close to shaving speed strokes. The setup can be adapted to several shaving systems to study their manipulation features. The project also provides us with knowledge on the current HSM module design and its possible extension in the future.

3

CONTENTS

Abstract ... 1

1. Introduction ... 5

1.1 Assessing shaving performance ... 5

1.1.1 Hair – Skin Manipulation (HSM)... 5

1.2 Goal of this Project ... 6

1.3 What are the core requirements? ... 6

1.3.1 Challenges of the project... 7

1.4 Available Choices ... 7

1.5 Stereo Imaging for 3D solution – what and why? ... 8

2. 3D Optical Setup ... 9

2.1 Camera selection and optical elements – High speed imaging ... 9

2.1.1 Numerical Aperture Conundrum ... 10

2.2 Illumination source selection – to be or not be bright ... 11

2.3 Mechanics – shaving motion made possible ... 11

2.4 Video Processing – Is software better than human eyes? ... 13

3. In – house built HSM Concepts ... 15

3.1 HSM principles used - how difficult is it to pull hairs? ... 15

4. Experimental Design ... 17

5. Main Results ... 19

5.1 Overall Performance ... 19

5.2 Observations – what went well and what did not ... 20

6. Conclusions ... 21

7. Future Work ... 23

5

1. INTRODUCTION

1.1 ASSESSING SHAVING PERFORMANCE

Hair removal, both for men and women, is a large and growing market. Philips is the market leader in the men’s electric shaving market. Furthermore, it offers a wide variety of products such as trimmers, mechanical- and light-based epilators [1].

Every hair removal device, such as shavers and epilators, works on a principle of hair and skin manipulation (HSM) which prescribes that the hair or skin or both are treated in a particular way such that the hair removal process becomes quick, painless, even and leaves the skin smooth with minimal stubble and minimal irritation. The essence lies is bringing the perfect balance between comfort and closeness. So far manual perception of closeness has been the best indicator of how well the shaving process has been done.

1.1.1 HAIR – SKIN MANIPULATION (HSM)

Fig.1.1: Hair- Skin manipulation concepts

Hair – Skin manipulation (HSM) refers to the techniques applied by a shaving system to manipulate the hair and skin such that hair can be removed easily with maximum efficiency. Fig.1.1 shows some of the prescribed techniques of HSM. While the first concept (Retraction) shows the manipulation of hair, the next two concepts show the manipulation of skin which would then indirectly manipulate hair. In general, shaving systems use a combination of both hair and skin manipulation.

In order for a HSM process to be successful, the hairs need to be pulled out from skin and placed in a position which is ideal for cutting. In general terms this pulling out of hair is called ‘hair- lift’ and placing of hair at ideal position suitable for cutting is called as ‘catching’. The extent of hair lift and the synchronization of hair lift with catching occurrence in a smooth (minimal irritation) way would describe the success of the HSM process.

Leaving aside the ‘smooth’ descriptor, it is important to quantify then the hair lift and the catching occurrence to describe the efficiency of the HSM process. This is where an optical system is required which would help in obtaining 3D information about the beard hairs undergoing HSM processes.

6

1.2 GOAL OF THIS PROJECT

The end goal of the project is to make a better shaving system which can give a close and smooth shave with minimal irritation. The goal of this project is to quantify or obtain the 3D information of the hair – skin manipulation process in men’s shaving system. Even though Philips has created several generations of electric shavers, there is still not a system which can easily quantify the efficiency of any HSM process. The quantification here refers to the measurement of 3 key parameters:

1. Change in the length of hair due to manipulation. 2. Change in the angle of hair due to manipulation.

3. Hair catching efficacy of each of the manipulation concept.

Using these key parameters it is possible to compare the performance of the existing shaving systems with the in-house developed HSM manipulation concepts. To be able to measure and to quantify the above mentioned parameters it is essential to have:

1. An optical system to visualize HSM process

2. Understand hair/skin manipulation methods and develop new concepts with improved performance with respect to market HSM performance indicators.

1.3 WHAT ARE THE CORE REQUIREMENTS?

Having explained the need to build a 3D optical setup, the core specifications for building such a system need to be decided. The system built should be such that it can clearly visualize beards (approximately 24h stubbles) with sufficient magnification suitable for high accuracy calculations, while the hair and skin are being manipulated at near to actual shaving speeds (up to 10cm/sec). It becomes very tricky to visualize hairs while it is being manipulated at such speeds without being blocked by physical components.

After considering, various aspects the core specifications of the 3D optical setup are given as: 1. Resolution: 10µm

2. Depth of focus: 300µm 3. Imaging speed : 1000-2000fps 4. Field of view: 1mm2 – 3mm2

The explanation for these specifications is explained as follows. The resolution is fixed around 10µm because the hair – lift or how far the hair comes out of the skin during the manipulation process when compared to its original state is what determines the closeness of the shave. In order to perform better than existing shavers, this parameter needs to be closely monitored down to stubble lengths in microns. A typical frequency of men shaving is once per 24h. Therefore, we need to visualize 24h hairs (~300 µm length). Therefore, the depth of focus has to be around 300µm. This way the image of the entire hair can be focused in with detail and clarity. In order to understand the hair – skin dynamics of a typical shaving process, we need to study it at the speeds at which shaving is done. This is estimated to be about 3 to 10cm/sec. This speed of shaving requires visualization speed at high frame rates of about

7

2000 – 3000fps. The field of view was chosen such that at least 1-3 hairs can be viewed with a good magnification at a time.

1.3.1 CHALLENGES OF THE PROJECT

1. A light weighted optical setup which satisfies the core specifications and which can be easily moved up and down to follow the shaving process across the face.

2. The optical setup should not be in contact with the face as then it would disturb the manipulation concept.

3. Viewing the manipulation concept through a tiny window, eg corresponding to the size of the blade slot (which is around 360 micrometers) while it occurs.

4. Extraction of 3D data from a camera giving out 2D information about the hair to determine the key performance parameters.

5. Proper illumination power, enough to achieve required frame rate and achieve the best contrast between hair and skin while avoiding artifacts such as hair visible under skin, reflections from skin etc.

6. A functional and smooth mechanical system on which the optical system can be placed for doing easy, safe and close to realistic shaving motions on volunteers without causing jitters to the optical system.

7. Automated software for analyzing the data received over hundreds of hairs in an accurate and quick manner.

8. The setup and the measurement technique should not be harmful or cause unacceptable stress to the volunteers.

1.4 AVAILABLE CHOICES

There are four popular methods through which 3D data can be extracted from an object [2]. The first technique is by using laser scanning where a narrow laser beam scans an object. A video camera captures the laser beam reflected by the surface and a computer further calculates the displacement of the reflected beam from the object surface against a straight line. With this the object surface can be obtained [3]. This method is robust and accurate but cannot be used for fast moving objects.

The second technique is light projection techniques which involve techniques like phase shift, colour or variable gratings projection [4]. This technique has grid patterns projected on the object and by extracting the grid or phase deformation of the reflected image, the surface profile of the object can be obtained. This method is faster than laser scanning but its accuracy is lower.

The third method known as time of flight method is popularly used in remote sensing [5]. Here a short pulse laser is used to measure the time of flight between the laser source and the object. This is a very expensive technique.

The final technique is called stereo imaging where two cameras placed at an angle between them are used to visualize the object of interest [6]. The disparity in 2D information obtained from both the cameras is used to obtain 3D information of the object. This technique is accurate and can be used to image objects in motion but requires accurate calibration and complex computation techniques.

8

1.5 STEREO IMAGING FOR 3D SOLUTION – WHAT AND WHY?

Considering the specifications and challenges it was proposed to build a stereo imaging system. Fig. 1.2 (left) shows an illustration of stereo imaging, as it occurs in the eyes and the way human eyes perceive objects in 3D.

Fig. 1.2: Stereo imaging in our eyes (left); Side and top views of beard in stereo imaging based setup (right)

Our two eyes are placed apart from each other at a distance of about 6 to 7 cm and they subtend an angle of 10 to 12 deg with each other. The left and the right eye each views a slightly different image which is called as a parallax. The brain takes in both the images and perceives them as one along with information of depth [7]. The same principle is applied in stereomicroscopes as well. There are two objectives and two eyepieces which magnify the image and help us see things in 3D. This is the concept used in the visualizing hair-skin manipulation in this project.

One of the previous works carried out at Philips involves using the principle of stereo-imaging. One of the images taken using a concept developed in-house for beard imaging is shown in Fig.1.2 (right). This previous system uses a single camera with a mirror to image the side and top view of the hair.

This setup is not suitable for hairs of all orientations and cannot visualize at actual shaving speeds. It also touched the face, which could unintentionally manipulate hairs.

Stereo imaging is ideal for our application as along with obtaining 3D data of the hair being manipulated, a very good qualitative assessment can also be made while analyzing various concepts. Thus in terms of accuracy of measurements, high speed imaging, safety of the volunteers and the advantage of obtaining qualitative analysis as well, stereo imaging was decided as the working principle of our optical setup. Appendix A gives the mathematical proof around the building of the optical setup.

9

2. 3D OPTICAL SETUP

As discussed in chapter 1, the goal of the project is to build a 3D setup which can accurately image the hair-skin manipulation (HSM) process and estimate from it the hair lengths and angles that they make with the skin. Based on our core specifications, stereo imaging was chosen as the design principle. On this foundation, this chapter goes into the design and construction of the 3D opto-mechanical imaging system.

Fig. 2.1 shows a simple illustration of how the 3D optical setup is designed. Here instead of the two eyes, two cameras are placed at an angle θ with respect to the normal. The illumination source is placed in between the two cameras. Appropriate lenses are attached to the cameras to meet the requirements. The two cameras image exactly the same area while the HSM process is carried out using a HSM module in contact with a volunteer’s face.

Fig. 2.1: Illustration of 3D setup

A high-speed, high resolution stereo imaging system includes the following key components: 1. Optical components that satisfy the required spatial resolution and field of view.

2. Illumination to achieve the required frame rate and obtain a good contrast between hair and skin.

3. Mechanics of the camera system and of the hair-skin manipulation module that can simulate actual shaving process.

4. Video Processing done by automated software which would then provide 3D information on the stubble length, angle and hair catching efficiency.

2.1 CAMERA SELECTION AND OPTICAL ELEMENTS – HIGH SPEED IMAGING

The camera selection is a critical part of building the setup and all other parts of the setup revolve around the selection of the cameras. The main selection criteria for these cameras were as follows:

1. Cameras that can image at high speed (~ 2000fps). Normal videos (seen on television and cameras) image at 25 – 30fps, and

10

2. Cameras that are light weight so that they can be moved at shaving speeds of up to 10cm/sec. Once the cameras were selected, desired optical elements were added to the setup such that the core specifications were met.

2.1.1 NUMERICAL APERTURE CONUNDRUM

Among several factors, selection of the numerical aperture (NA) of the camera lenses was an interesting choice. Numerical aperture is the light gathering power of the system given as [8]:

Where n is the refractive index of the medium (n = 1, for air) and θ is the half angle of the acceptance cone formed by the lens as shown in Fig.2.2. D is the diameter of the entrance pupil if the lens and f is the focal length of the lens.

Fig.2.2: Defining numerical aperture (left); final image obtained from camera (right)

The NA affects several other important parameters such as the resolution of the system, the depth of focus and frame rate achieved. As the NA increases, the resolution improves and the amount of light entering the lens increases. Taking movies at high frame rate requires that maximum amount of light is captured. On the other hand, a high NA means a low depth of focus. Thus an object of even small dimensions appears fuzzy in depth. By playing around with NA, we could fix the 3 most important requirements of the system.

As seen in Fig. 2.3, the image in the left is taken with a higher NA. The image has good depth of focus but is darker. The image on the right is taken with lower NA. It has poor depth of focus but is much brighter.

11

Fig. 2.3: Impact of NA on depth of focus: lower NA (left); higher NA (right)

Fig. 2.2 (right) shows a sample image taken using 2 cameras (angle of 40 deg between them) with a field of view of 2mm x 2mm, frame rate of 2000fps, resolution of 15μm and depth of focus of 500μm.

All the details surrounding the selection of cameras and associated optical elements are given in Appendix B.

2.2 ILLUMINATION SOURCE SELECTION – TO BE OR NOT BE BRIGHT

The decision of selecting an appropriate light source had the following major criteria: 1. Sufficient power to obtain the high frame rate movies and sufficient hair – skin contrast. 2. The light source and its attachments must fit comfortably between the two cameras.

3. Avoiding reflections which cause bright spots on hair and skin. This can cause confusion as to where the hair begins and ends.

4. Imaging only the hair above the skin. Images are created which show under the skin hair (Fig. 2.2, right). This can also cause confusion as to where exactly the hair begins.

All the experiments were done using a halogen lamp source with a power of 150W. The source is brought towards the object (hair-skin) with a flexible fibre guide. Appendix C gives the overview over the decision of the choices made during the selection of light source.

2.3 MECHANICS – SHAVING MOTION MADE POSSIBLE

The first setup involved a table top arrangement, with the cameras and the lighting attached to the table and with the person sitting in front of the HSM module and moving himself up and down to manipulate hair and skin (Fig. 2.4). This was hardly close to actual shaving conditions. Eventually the setup evolved into a device as shown in Fig.2.5.

12

Fig.2.4 Initial setup

The setup was constructed with several degrees of freedom suitable for measuring comfortably in three regions – cheek, chin and neck. The setup also has a software controlled linear translation stage which on a button click moves up and down (along with cameras, HSM module and lighting) at shaving speeds. A force sensor is attached to the module to measure the normal force experineced by the module from the user while shaving.

Further information on the Mechanics of the setup are given in Appendix D.

13

2.4 VIDEO PROCESSING – IS SOFTWARE BETTER THAN HUMAN EYES?

The 3D optical setup is now ready to make measurements. One of the bottle necks at this juncture is to arrive at a technique which would help in easily analyzing such high frame rate movies. The initial idea was to develop a fully automatic software which would by itself assess data. This was tricky as the issue of where exactly the hair ends and begins was debatable even with manual inspection. Fig.2.6 shows one such troublesome hair to illustrate this problem. To leave this decision to the software proved error prone.

Fig. 2.6: Confusion in estimating hair’s true length

Fig. 2.7: Screenshot of a movie being processed by semi-automated software

Hence semi-automatic software was created on which movies are shown side by side by the software (Fig.2.7). The software allows for manual selection (clicking) of the beginning and end of the hair at each frame of the movie. Once the selection is made, it gives out the length, angles associated with the hair and several other measurements. Appendix E gives a short description of the software used. The entire 3D setup is calibrated and a detailed error analysis is done. Appendix F and G illustrate further on this.

15

3. IN – HOUSE BUILT HSM CONCEPTS

Fig. 3.1 HSM Module: Front view (left); side view (right)

Above Fig. 3.1 shows the first hair – skin manipulation (HSM) concepts to be used for a new generation shaving systems. The goal of the module is to manipulate hair and skin such that a close and smooth shaving occurs. This particular module has been specifically designed to manipulate “against the grain” hairs, which means the module attacks the bottom of the hair before it reaches its root. The yellow square is the actual area imaged by the two cameras. It is important to note here that the ‘blade’ here performs no cutting action. It is designed only to manipulate skin and hair.

A total of six different module geometries are made based on the blade exposure and slit size (gap between stretcher and the blade edge).

3.1 HSM PRINCIPLES USED - HOW DIFFICULT IS IT TO PULL HAIRS?

Fig. 3.2: Implemented major concepts in current HSM module

Fig. 3.2 (top) shows the various concepts that were implemented using the current module and (bottom) concepts are the ones that occur sometimes depending on the module geometry and the size of the hair. Appendix H gives more insight into the HSM principles used and how they are implemented in the software.

17

4. EXPERIMENTAL DESIGN

The 3D optical setup is built, calibrated and the HSM concepts are ready to be tested. Before beginning the experiments all the variables associated are assessed to get an idea of which parameters are essential and to what extent they can and should be controlled and monitored (Appendix I).

This section describes how the experiments are performed on a single hair. There are essentially three regions of interest – chin, cheek and neck. Beards on all these regions are chosen for experiments. A hair is chosen on a specific region and it is circled by a marker pen for clear identification. Every hair is visualized 3 times (Fig. 4.1):

1. Unmanipulated stroke: This video is taken by placing the subject in front of the camera and filming the hair without any manipulation done by the HSM module. The purpose of this experiment is to calculate the length of the hair without any manipulation done on it.

2. Dynamic stroke: This video is taken by placing the module below the target hair and making a stroke such that the module passes over the selected hair. The dynamic experiment is to simulate natural shaving process such that the selected hair is one of the hairs filmed during the manipulation process.

3. Static to dynamic stroke: In this case the module is placed such that the selected hair is in the slit and field of view to begin with. The movie is taken when the hair is in a static position and from that position the module moves on top of it. This way we can see what exactly how a hair behaves when it is placed in the slit. These experiments are done for our further understanding.

Fig.4.1: Unmanipulated (left); dynamic (middle); static to dynamic (right) strokes made on each hair

At each stage the lengths and angles related to the hairs can be evaluated by simply clicking on the bottom and top of the hair.

19

5. MAIN RESULTS

Using six module geometries experiments were performed on six volunteers on their cheek, chin and neck beard hairs (approximately 24h stubbles). A total of 40 hairs and 80 strokes (repeated 2twotimes for each hair) were tested upon.

5.1 OVERALL PERFORMANCE

The below results show a flavor of the type of results expected from the experiments performed.

Fig. 5.1: Overall Results, graph (top); column chart (bottom)

Fig.5.1 (top) shows the histogram of the lengths obtained of unmanipulated, manipulated (mentioned as ‘length at focus’) and remaining stubbles. The remaining stubble lengths are obtained assuming that there is an actual cutting element instead of the POM material on the blade and the hair is cut. There is a trend for the manipulated hairs (green line) to have longer hair lengths which indicates a good manipulation and hair lift and also a trend for remaining stubbles (purple line) to have shorter lengths. The column chart (Fig. 5.2, bottom) shows the average lift of the hairs to be 82µm. 63% of the total hairs are manipulated into a position suitable for cutting.

20

In terms of the success of the HSM process, creating hair lift alone is not sufficient. The hair must also be brought into a position suitable for cutting. This second requirement is described as manipulation efficiency.

5.2 OBSERVATIONS – WHAT WENT WELL AND WHAT DID NOT

From the experiments some of the conclusions obtained are:

1. The blade does not have sufficient adherence to the skin. This is the biggest problem faced by the current HSM concept. The blade seems to be ‘floating’ instead of making contact with skin in most strokes.

2. Low blade exposure (0µm) and small slit size (0.5mm) show poor hair manipulation efficiency as these modules do have sufficient blade exposure and also when the hair is long (with respect to slit size), it gets flipped under the rubber first and it does not have enough time to flip back into the region suitable for cutting.

3. An exposure of 220µm and a slit size of 1mm is the best module geometry. 80% of the hairs are manipulated.

4. Rubber stretcher has generally more effect than the blade irrespective of the exposure of the blade. The friction between the rubber stretcher and the skin is large and hence can lead towards the uneven attachment of the module, leaving the blade without contact to skin. 5. For medium hairs and large slit compared to the hair, the rubber stretcher first flips the hair

beyond the acceptable region suitable for cutting, then the hair has enough time to flips back in to the region, then the blade manipulates it efficiently.

6. Chin is the most troublesome area for manipulation. The region is uneven making it difficult to follow the contour. Cheek is the easiest to tackle region.

7. In the most efficiently manipulated case, 100µm lift seems to be enough to obtain short (virtual) stubbles.

8. Force range between 3 to 4N gives the best combination of hair lift and manipulation efficiency. Appendix K gives a detailed explanation of the obtained results

21

6. CONCLUSIONS

As shaving systems claim closer and closer shaves, it becomes essential to verify these claims on available shaving devices and also on in-house developed concepts. This project aimed to visualize and quantify the hair lengths and angles during the hair – skin manipulation process prior to the cutting action in shaving process. While previously built setups could also visualize hairs, their accuracy suffered and the measurements could not be done on all free standing hairs without manipulating the skin. Stereo imaging was chosen as the working principle and based on it an opto-mechanical setup was designed and constructed. Two high speed cameras (Photron Fastcam MC2) placed at an angle of 40 deg with respect to each other were chosen as the cameras of the system which could capture movies at a frame rate of 2000fps (512 x 512 pixel resolution). The optical system has a resolution of 15µm, a depth of focus of 500µm, a magnification of 2X with a field of view of 2mm. Thus the core specifications of the system were met. A white light source with a 150W halogen lamp (KL-1500 LCD) guided by flexible fiber was used as an illumination source for the whole setup. The cameras, illumination and the developed manipulation module were all placed on a linear translation stage which moves at speeds up to 10cm/sec. With the setup convenient measurements could be made on cheek, neck and chin while person sits in front of the module.

The system was calibrated by using a phantom hair-skin sample and the measurements taken by Optical Coherence Tomography microscope were used as a standard. A detailed error analysis was done on the 3D setup by a building a mathematical model around it. The error margin in the measurement of hair lengths is about 3%.

Semi-automatic software was developed exclusively for the project using Matlab. The software allows the user to click the top and the beginning of the hair through which the length and angle details of the hair can be obtained. Laser markings were made on the module so that it is possible for the software to obtain the exact location of the hair with respect to the blade.

Elaborate experiments were done on cheek, chin and neck using six geometry modules with various blade exposures and slit sizes. The module with an exposure of 220µm and a slit size of 1mm was identified as the best with 80% hair manipulation efficiency. This module had an average hair lift of 110µm and average remaining stubble of 90µm. Chin is identified as the toughest area to manipulate due to difficulty in contour following. Taking into account only the hairs manipulated with the most efficient conditions, the average stubble length was -36µm.

In 35% of the incident strokes, the normal force applied during dynamic stroke was in the range of 1 to 2N. Maximum efficiency in terms of manipulation (88%) and lift (80%) though are found to be between the force range of 3 to 4N.

23

7. FUTURE WORK

Several experiments were performed on beard hairs on cheek, chin and neck to check the efficiency of the in-house built manipulation module. Some of the suggestions leading towards future activities are:

1. Understanding success and failure mechanisms - Detailed understanding of the various mechanisms is needed to fully comprehend and further implement ideas to improve the manipulation efficiency. One way to do this could be by performing experiments with dummy modules containing only blade element, only the rubber stretcher with or without lubra strip etc., to understand the role played by each element in the module.

2. Contour following - In some of the experiments it was observed that the blade is not in contact with the skin. One of the reasons for this could be poor contour following by our module. The setup has to be changed to accommodate better following of the face contour, especially the chin.

3. Module designs - Having obtained knowledge over the current module geometries’ functionalities it would be interesting now to perform experiments with modifications on existing designs (larger exposure, smaller rubber size etc.) as well as new designs.

4. Role of skin friction - Experiments were performed with shaving gel on each volunteer. In future experiments the amount of gel used and the gap between each application must be controlled as the skin friction coefficient increases with time after application of these gels [9,10] making it counterproductive towards easier manipulation.

5. Finite element modeling - Modeling may help in creating a better understanding the dynamics of hair and skin.

6. Benchmarking - At this stage it is critical to check the performance with a market benchmark in terms of shaving. Thus experiments need to be performed to set such a benchmark. This could involve adjusting the setup to possibly view such a shaving process.

25

REFERENCES

1. Philips. ‘Annual Report 2009’, [available from - http://www.annualreport2009.philips.com/]

2. T. T. Lu and T.H Chao, A high-resolution and high-speed 3D imaging system and its application on ATR, Proc. SPIE 6245, SPIE Symposium on Defense and Security, Optical Pattern Recognition XVII, 2006

3. M. Levoy, J. Ginsberg, J. Shade, D. Fulk, K. Pulli, B. Curless, S. Rusinkiewicz, D. Koller, L. Pereira, M. Ginzton, S Anderson, J. Davis, "The Digital Michelangelo Project: 3D Scanning of Large Statues," Proceedings of the 27th annual conference on Computer graphics and interactive techniques, pp.131-144, 2000.

4. G. J. Agin and P. T. Highnam, "Movable Light-Stripe Sensor for Obtaining Three-Dimensional Coordinate Measurements," SPIE Proc. Vol. 360, p. 326, 1983.

5. N. Haala, and C. Brenner, “Generation of 3D city models from airbrone laser scanning data,” Proceedings ofEARSEL workshop on LIDAR remotesensing on land adn sea, Tallin, Estonia, pp. 105-112, 1997.

6. R. Koch, M. Pollefeys and L. Van Gool, Automatic 3D Model Acquisition from Uncalibrated Image Sequences, Proceedings Computer Graphics International, pp.597-604, Hannover, 1998

7. ‘Stereo Vision’, [available from - http://www.vcbio.science.ru.nl/en/image-gallery/stereo/] 8. E.Hecht, ‘Optics’, 4th edition, Addison Wesley Longman, Inc. 1998

9. R.K. Sivamani, J.Goodman, N.V.Gitis, H.I Maibach, ‘ Fricton coefficient of skin in real time’, Skin Research and Technology, 9, 235-239, 2003

10.M. J. Adams, B. J. Briscoe, S. A. Johnson,’ Friction and lubrication of human skin’, Tribology Letters, Vol. 26, No. 3, 2007.