Hairware: Conductive Hair Extensions as a Capacitive Touch Input Device

Katia Vega Pontifical Catholic University of

Rio de Janeiro Rua Marquês de São Vicente, 225 - Gávea, Rio de Janeiro -

RJ, Brazil kvega@inf.puc-rio.br

Marcio Cunha Pontifical Catholic University of

Rio de Janeiro Rua Marquês de São Vicente 225 – Gávea, Rio de Janeiro –

RJ, Brazil mcunha@inf.puc-rio.br

Hugo Fuks Pontifical Catholic University of

Rio de Janeiro Rua Marquês de São Vicente, 225 - Gávea, Rio de Janeiro -

RJ, Brazil hugo@inf.puc-rio.br

Figure 1 – Designing Hairware as an input device. a) Changing capacitance by a three layers approach. b) Hair extensions during chemical process. c) Visual feedback while touching the hair.

ABSTRACT Our aim is to use our own bodies as an interactive platform. We are trying to move away from traditional wearable devices worn on clothes and accessories where gestures are noticeable and remind cyborg looking. We follow Beauty Technology paradigm that uses the body’s surface as an interactive platform by integrating technology into beauty products applied directly to one’s skin, fingernails and hair. Thus, we propose Hairware, a Beauty Technology Prototype that connects chemically metalized hair extensions to a microcontroller turning it into an input device for triggering different objects. Hairware acts as a capacitive touch sensor that detects touch variations on hair

and uses machine learning algorithms in order to recognize user’s intention. In this way, we add a new functionality to hair extensions, becoming a seamless device that recognizes auto-contact behaviors that no observers would identify. This work presents the design of Hairware’s hardware and software implementation. In this demo, we show Hairware acting as a controller for smartphones and computers.

Author Keywords Hairware; Conductive Hair; Beauty Technology; Wearable Computing; gesture recognition.

ACM Classification Keywords H.5.2 Information interfaces and presentation: User Interfaces; Haptic I/O. INTRODUCTION When embedding technology into everyday life objects, Weiser envisioned it functioning in an invisible and unobtrusively way [14]. For that reason, minimizing awareness while interacting with technology, continue to be an important guideline in Ubiquitous Computing. Our concern with this vision is that users are not consciously aware that these pieces of technology are gathering data and processing it for them. In contrast, we propose interactions

Permission to make digital or hard copies of part or all of this work for personal or classroom use is granted without fee provided that copies are not made or distributed for profit or commercial advantage and that copies bear this notice and the full citation on the first page. Copyrights for third-party components of this work must be honored. For all other uses, contact the Owner/Author. Copyright is held by the owner/author(s). IUI'15 Companion, Mar 29 - Apr 01, 2015, Atlanta, GA, USA ACM 978-1-4503-3308-5/15/03. http://dx.doi.org/10.1145/2732158.2732176

Non–conductive hair extensions

Layer 1

Layer 2

Layer 3

Conductive Hair

a) b) c)

that the user is consciously aware of and intends to generate them, though in a concealed way, by enacting auto-contact behaviors.

Even though the billionaire beauty business market is mostly supported by women and most women wear makeup everyday [4], beauty products have not yet been thoroughly explored in relation to their use as wearable computing. Beauty Technology [10,11,12] disrupts this frontier by placing technology directly on the body surface allowing women to use beauty products to control their environment through technology in a personal, seamless and fashionable way.

This work proposes the use of Hairware as a novel electronic device to be used in wearable computing. We use a chemical platting technique that makes the artificial hair extensions conductive but, at the same time, remains looking like regular human hair. This paper introduces Hairware, a capacitive touch sensor that detects a variety of touches for the triggering of different devices. Section 2 identifies previous work on body technologies as input devices. Section 3 presents the hardware and software implementation to detect touch on Hairware. Section 4 shows the model used for learning and recognizing touches on hair. Section 5 presents the evaluation results and reviews the lessons learned using Hairware. Conclusion and future work are shown in the last section.

RELATED WORK Wearable computing changed the way individuals interact with computers, intertwining natural capabilities of the human body with processing apparatus [14]. Most of this technology has been designed just for clothing or accessories. However, some works place technology on the body’s surface. Thus, the skin and its regenerative appendages (i.e. hair and nails) not just play a crucial role in human beings but also embeds electronics to act as input devices. For example, Skinput [2] uses a bioacoustics microphone to sense the vibration that occurs when the body is tapped. Another example is Senskin [5], a wearable device that uses multiple IR reflective sensors implanted in an armband to allow the skin to be used as a soft input interface by comparing the actual force and the sensor values using an elastic model of the human skin. Touché [6] is a sensing technology that proposes a Swept Frequency Capacitive Sensing technique by using bracelets that detects touches on different parts of the body by sensing the impedance variability between electrodes. Thus, the user could put the phone into silent mode by making a “shih” gesture bringing her index finger to the lips.

In previous Beauty Technology prototypes, Conductive Makeup [8,9] was presented as an interface for detecting voluntary blinking. Conductive Makeup includes conductive eyeliner and chemically metalized fake eyelashes that act as blinking switches. While conductive eyeliners connect sensors and actuators by using conductive

materials that stick to the skin, replacing conventional eyeliners, conductive fake eyelashes sense blinking.

Wigs could be used either as apparel or to fulfill cultural or religious obligations. SmartWig [9] is a wearable device that uses a wig for hiding electronics that communicates wirelessly with other external devices. SmartWig suggests applications that could fulfill a number of functions, from acting as a health care device that monitors users’ vital signs to helping blind people navigate roads, or changes slides in a presentation by tapping their sideburns, under which buttons are hidden. A further potential improvement of the wig may use ultrasound waves to detect objects around a user. Hair accessories that vary from clips to corsages, designed to give every look a final flourish, could also be used for creating discreet and fashionable gadgets attached to the hair. First Sign Hair Clip [8] is an electronic hair clip that communicates with a mobile application to automatically call for help and collect evidence when the user is in danger. By using accelerometers and gyroscopes, it detects head impacts associated with violent crimes, automatically triggering an alarm, while evidence is collected by a microphone.

DESIGNING HAIRWARE Our prototype combines hardware (hair extensions, circuit and microcontroller) and software (machine learning algorithm).

Hardware Hairware is artificial hair extensions that are chemically metalized for acquiring electrical conductivity and keeping a natural coloration. The chemical process is carried out in two phases: Activation and Electrolysis (Figure 1b). We used a Hairware strand of approximate 1.5 by 25cm and with a surface resistivity of less than 5 ohm/sq.

Hairware is connected to a circuit to work as a capacitor sensor. This circuit creates a delay in the pulse that is the time the capacitor takes to charge and discharge. In this way, Hairware is used as a conductive surface that detects when another conductive surface approximates to it. Therefore, as the human body is conductive, the average internal resistance of a human trunk is ~100 Ω [13], touching Hairware will affect its capacitance and result in a different charging time.

We use hair clips for attaching the circuit to the hair extensions in order to make it easily removable and replaceable. A wire connects the conductive hair extensions to the microcontroller, in this way, the conductive hair extension keeps rounded by the non-conductive hair and are not directly in contact with the user’s nape. In addition, this approach makes it possible to put the circuit in different accessories such as headbands, brooches, hats and the top of the hair extensions.

Figure 1a shows three layers of non-conductive hair extensions added for isolating the hair from the skin and nape. In addition, these layers improve the capacitor sensor

values. Each time the user touches the top, middle or tip, the capacitor sensor differentiates these values. The circuit compares an output that transmits the pulse and an input, which receives the pulse. When a finger touches Hairware, it creates a delay in the pulse, and the Arduino microcontroller recalculates this delay. The circuit has four 1MΩ resistors and one 100pF capacitor. The resistors selects the sensitivity, bigger the resistor, the farther away it detects a human. With 4MΩ resistor between the output and input pins, the circuit is tuned to start to respond one inch away, just the sufficient to overcome the non-conductive hair layer. The small capacitor (100 pF) placed from sensor pin to ground improves stability and repeatability. Some LEDs were added to the system to give feedback to the user whenever a touch is detected (Figure 1c).

Software For classification, we use a decision tree implementation provided by BigML [1]. Each transmission from the sensor contains a capacitive charging time in milliseconds, which we extract for classification. Currently, we can classify five touches interaction in the Hairware: no touch, touch on the top, touch on the middle, touch on tip and straightening.

When trained, the decision tree build an actionable model that we use as an Arduino function to classify the interaction in real time. Each node represents the confidence action from the read value. The key feature to recognize touch on the hair is not to detect the touch itself, but the circuit must be fined tune to detect small capacitance charge variation along the hair.

As expected, the touch on the tip has the higher value due to the fact that layer of non-conductive hair is the thinner one (layer 3). The charging value will decrease when the interaction gets near the top where there are more non-conductive hair layers acting as a filter. The straightening action has the lowest values due to its continuous contact time with the sensor. Noise appeared on the readings given that the hair was in direct contact with the skin. However, our algorithm recognizes it as noise and obtains 92% accuracy on the five trained touch gestures.

EVALUATION For our evaluation, we aim to identify the exposure of the technology in Hairware and we measure the accuracy in replaying the hair straightening, top, middle, and tip touch. We recruited five female participants, ranging in age from 28 to 35. The study took approximately 30 minutes per each participant and included a gratuity. Given that Hairware is a proof of concept of the technology, evaluations were taken in the same laboratory under the same conditions (temperature, humidity and pressure). At the beginning, we conducted pre-questionnaires to identify the participants’ profiles. All participants were right handed. They do not own any wearable technologies but every day they use other electronic devices like smartphones and tablets. None had used hair extensions before but they all knew what they were and they have seen them applied to other person. All

of them expressed that they touch their hair when they feel anxious, thoughtful and nervous. Three of them noticed them using their hair as a flirting tool.

We interview the participants before they observed and used the wearable device. We asked them questions related to Hairware visibility and their use of hair extensions. While showing Hairware, we asked them about their impressions about it, if they could identify any difference between these hair extensions, and normal or artificial hair. Then we showed a non-conductive artificial hair extension and asked about the differences they could detect between this and Hairware.

After 2 minutes of training on the different tasks, we asked them to perform the five different touches. They performed each touch 10 times. Finally, we conducted an interview asking for recommendations and future uses.

RESULTS AND DISCUSSION After listening to the answers of the first interview, we identified that, at first sight, Hairware wasn’t noticed as a technological device and they weren’t aware of its conductivity. Most of the participants thought it was natural hair extensions and comparing it to artificial hair, they noticed that the artificial hair extension was more shining and had a smoother texture. They also expressed that, when Hairware was attached to someone’s hair, the difference between it and natural hair was not noticeable.

During the tests, we observed that, although our algorithm differentiated each touch, a previous calibration was necessary for each user. Differences in skin capacitance and hand position could made the capacitor sensor identify different values for each user. We also noticed that other feedback must be designed. LEDs won’t be noticed if the user is wearing Hairware. We will replace this feedback by vibration motors.

After using Hairware, they expressed that adding layers of non-conductive artificial hair extensions, gave the device a shiny look and it was smooth to touch. They also expressed that, if the circuit is hidden into the hair or into an accessory, an external observer won’t notice the device. Future versions of Hairware will add other touches that the participants exposed that they usually perform like twirling their hair and passing the fingers through the head.

During the final interview, participants expressed that they would use Hairware for specific purposes like a security device. That is, whenever they feel in danger, they would like to send a message to the police or to someone they trust without anyone around noticing that they were doing it. Another suggestion was to use it as an input device for triggering house appliances. Another suggestion was to trigger a reminder whenever they are continuously touching their hair as a nervous tick. In addition, it could be used as a therapy tool for relaxing when the device noticed they are stressed.

CONCLUSION AND FUTURE WORK Hairware fosters a seamlessly looking approach to wearables. It is artificial hair extensions that are chemically metalized to maintain a natural coloration and when connected to a microcontroller could be used as an input device. Thus, not just the technology is seamless but also the action of touching her hair that trigger devices are hardly noticed by an external observer. This paper explained our processes in designing the interface, creating the hardware and implementing the algorithm for detecting touches on the top, middle, tip, straightening the hair and no touch. A study case was conducted with five participants to identify if the technology is noticeable by an observer and for testing our gesture recognition system. Possible uses for it would be any kind of communication without being noticed. A security device could trigger messages with the user’s location and recording the event. Customized actions on the phone could make a call, record an audio, send a default message and take a picture. Artistic performers could use Hairware to turning on lights and devices on scene. Also secret agents could start filming and communicate with other agents.

Future works will include the use of beard Hairware in order to explore the masculinity side of the equation. In addition, it will be used as both, input and output devices. Actuators as LEDs, speakers and vibration motors could be used to explore Hairware as an output device. Another input interface could be an array of involuntary acts to sense and control anxiety and stress in users.

Other sensors, such as accelerometers and gyroscopes will be integrated on the hair extensions to make combinations of movements and hair touch. A potential improvement may be to use ultrasound waves to detect objects around a user [3]. Other movements like curling Hairware with fingers may be explored when using Hairware as flex sensors [7].

ACKNOWLEDGMENTS Marcio Cunha (grant 163066/2013-2) and Hugo Fuks (grant 302230/2008-4) are recipients of grants awarded by the National Research Council (CNPq). This work was partially financed by Research Support Foundation of the State of Rio de Janeiro-FAPERJ/INCT (E-26/170028/2008) and CNPq/INCT (557.128/2009-9). Katia Vega is a postdoctoral fellow with grant funding from PNPD/CAPES Portaria 086/2013.

REFERENCES 1. BigML. BigML. https://bigml.com/. 2. Harrison, C., Tan, D., and Morris, D. Skinput:

appropriating the body as an input surface. Proceedings of the SIGCHI Conference on Human Factors in Computing Systems, (2010), 453–462.

3. Kasahara, S., Nagai, S., and Rekimoto, J. LiveSphere: Immersive Experience Sharing with 360 Degrees Head-mounted Cameras. Proceedings of the Adjunct Publication of the 27th Annual ACM Symposium on User Interface Software and Technology, ACM (2014), 61–62.

4. Mint. Beauty or Bust. Obsessed with cosmetics. https://www.mint.com/blog/consumer-iq/splurge-vs-save-which-beauty-products-are-worth-the-extra-cost-0413/?display=wide.

5. Ogata, M., Sugiura, Y., Makino, Y., Inami, M., and Imai, M. Senskin: adapting skin as a soft interface. Proceedings of the 26th annual ACM symposium on User interface software and technology, (2013), 539–544.

6. Sato, M., Poupyrev, I., and Harrison, C. Touch{é}: enhancing touch interaction on humans, screens, liquids, and everyday objects. Proceedings of the SIGCHI Conference on Human Factors in Computing Systems, (2012), 483–492.

7. Schoessler, P., Leigh, S., Jagannath, K., van Hoof, P., and Ishii, H. Cord UIs: Controlling Devices with Augmented Cables. Proceedings of the Ninth International Conference on Tangible, Embedded, and Embodied Interaction, ACM (2015), 395–398.

8. Sign, F. Products - First Sign. 2014. http://www.firstsign.us/products/.

9. Tobita, H. and Kuzi, T. SmartWig: wig-based wearable computing device for communication and entertainment. Proceedings of the International Working Conference on Advanced Visual Interfaces, (2012), 299–302.

10. Vega, K. and Fuks, H. Beauty Technology: Body Surface Computing. Computer 47, 4 (2014), 71–75.

11. Vega, K. and Fuks, H. Beauty Technology As an Interactive Computing Platform. Proceedings of the 2013 ACM International Conference on Interactive Tabletops and Surfaces, ACM (2013), 357–360.

12. Vega, K. and Fuks, H. Empowering Electronic Divas through Beauty Technology. Lecture Notes in Computer Science, Springer (2013), 237–245.

13. Webster, J.G. Medical instrumentation. Application and Design, Houghton Mifflin Company, Boston. 197& g, (1973).

14. Weiser, M. The computer for the 21st century. Scientific american 265, 3 (1991), 94–104.