The pottery skills and tacit knowledge of a maser: An analysis using eye-tracking data

Abstract

An important aspect of Japanese traditional culture is the design and creation of pots, bowls, and other pottery, an artform that has historically been subject to cultural influences from China and the Korean peninsula. It has been suggested that passing down of hand-crafting techniques to the next generation is quite difficult, as the techniques involve considerable tacit knowledge.
Thus, this study aimed to explore the creative processes of a master of the craft using an eye-tracking system. We recorded the eye movements of Hakuga Nakamori, a third-generation artisan, as he created a bowl and a pot. The results showed that, in the process of working toward creating a piece of art, Mr. Nakamori first focused his attention on the whole piece, and then on certain elements thereof.

Full text

ScienceDirect
Available online at www.sciencedirect.com
Procedia Computer Science 159 (2019) 1680–1687
1877-0509 © 2019 The Authors. Published by Elsevier B.V.
This is an open access article under the CC BY-NC-ND license (https://creativecommons.org/licenses/by-nc-nd/4.0/)
Peer-review under responsibility of KES International.
10.1016/j.procs.2019.09.338
10.1016/j.procs.2019.09.338 1877-0509
© 2019 The Authors. Published by Elsevier B.V.
This is an open access article under the CC BY-NC-ND license (https://creativecommons.org/licenses/by-nc-nd/4.0/)
Peer-review under responsibility of KES International.
Available online at www.sciencedirect.com
ScienceDirect
Procedia Computer Science 00 (2019) 000–000
www.elsevier.com/locate/procedia
1877-0509 © 2019 The Author(s). Published by Elsevier B.V.
This is an open access article under the CC BY-NC-ND licen se (https://creativecommons.org/licenses/by-nc-nd/4.0/)
Peer review is the responsibility of KES International.
23rd International Conference on Knowledge-Based and Intelligent Information & Engineering
Systems
The pottery skills and tacit knowledge of a maser:
An analysis using eye-tracking data
Jun Nakamuraa*, Sanetake Nagayoshib
aChuo University, 742-1 Higashinakano, Hachio-ji, Tokyo 192-0393, Japan
bShizuoka University,
3-5-1 Johoku, Naka-ku, Hamamatsu, Shizuoka 432-8011
, Japan
Abstract
An important aspect of Japanese traditional culture is the design and creation of pots,bowls, and other pottery, an artform that has
historically been subject to cultural influences from China and the Korean peninsula. It has been suggested that passing down of
hand-crafting techniquesto the next generation is quite difficult, as the techniquesinvolve considerable tacit knowledge.
Thus, this study aimed to explore the creative processes of amaster of the craft using an eye-tracking system. We recorded the eye
movements of Hakuga Nakamori, a third-generation artisan, as he created a bowl and a pot.The results showed that, in the process
of working toward creating apiece of art, Mr.Nakamori first focused his attention on the whole piece,and then on certain elements
thereof.
© 2019 The Author(s). Published by Elsevier B.V.
This is an open access article under the CC BY-NC-ND license (https://creativecommons.org/licenses/by-nc-nd/4.0/)
Peer review is the responsibility of KES International.
Keywords: Eye tracking; cognitive process; moulding
1. Introduction
Japan has the longest history of pottery-making in the world, although the development of this artform has been
subject to cultural influences from China and the Korean peninsula [1]. The oldest pottery pieces in the world, which
are about 12,000 years old, were excavated in Japan. During the 4th to mid-5th centuries,the potter’s wheel, which
enabled artisans to create works of various shapes, was introduced to Japan from the Korean peninsula. Later, the kiln
* Corresponding author.
E-mail address: jyulis.77f@g.chuo-u.ac.jp, jyulis@tamacc.chuo-u.ac.jp
Available online at www.sciencedirect.com
ScienceDirect
Procedia Computer Science 00 (2019) 000–000
www.elsevier.com/locate/procedia
1877-0509 © 2019 The Author(s). Published by Elsevier B.V.
This is an open access article under the CC BY-NC-ND licen se (https://creativecommons.org/licenses/by-nc-nd/4.0/)
Peer review is the responsibility of KES International.
23rd International Conference on Knowledge-Based and Intelligent Information & Engineering
Systems
The pottery skills and tacit knowledge of a maser:
An analysis using eye-tracking data
Jun Nakamuraa*, Sanetake Nagayoshib
aChuo University, 742-1 Higashinakano, Hachio-ji, Tokyo 192-0393, Japan
bShizuoka University,
3-5-1 Johoku, Naka-ku, Hamamatsu, Shizuoka 432-8011
, Japan
Abstract
An important aspect of Japanese traditional culture is the design and creation of pots,bowls, and other pottery, an artform that has
historically been subject to cultural influences from China and the Korean peninsula. It has been suggested that passing down of
hand-crafting techniquesto the next generation is quite difficult, as the techniquesinvolve considerable tacit knowledge.
Thus, this study aimed to explore the creative processes of amaster of the craft using an eye-tracking system. We recorded the eye
movements of Hakuga Nakamori, a third-generation artisan, as he created a bowl and a pot.The results showed that, in the process
of working toward creating apiece of art, Mr.Nakamori first focused his attention on the whole piece,and then on certain elements
thereof.
© 2019 The Author(s). Published by Elsevier B.V.
This is an open access article under the CC BY-NC-ND license (https://creativecommons.org/licenses/by-nc-nd/4.0/)
Peer review is the responsibility of KES International.
Keywords: Eye tracking; cognitive process; moulding
1. Introduction
Japan has the longest history of pottery-making in the world, although the development of this artform has been
subject to cultural influences from China and the Korean peninsula [1]. The oldest pottery pieces in the world, which
are about 12,000 years old, were excavated in Japan. During the 4th to mid-5th centuries,the potter’s wheel, which
enabled artisans to create works of various shapes, was introduced to Japan from the Korean peninsula. Later, the kiln
* Corresponding author.
E-mail address: jyulis.77f@g.chuo-u.ac.jp, jyulis@tamacc.chuo-u.ac.jp
2Author name / Procedia Computer Science 00 (2019) 000–000
was introduced,providing a sophisticated system for heating pottery to 1,000°C. This solved the problem of water
leakage,such that the pottery showed durability to liquids.
As an expression of traditional Japanese culture,pottery making obviously requires old-fashioned hand-crafting
techniques,despite recent developments in digital technology.Some have suggested that it is very difficult to pass
such techniques on to the next generation [2,3]. This study aims to draw attention to this artform to avoid the loss of
pottery-making techniques that are based on tacit knowledge [4], and to preservethe traditional skillsfor future
generations. Verbalising the techniques by which these skills could be conveyed is difficult [5], so this study focused
on eye movements. When recognising objects, people unconsciously come to an understanding of the ‘what’ of the
object; asthe proverb goes, “The eyes are the windows to the soul”. Another old proverb suggests that one should
“learn by seeing his back”, i.e. that learners should acquire skills by observing veterans. For an example, Japanese
martial arts like judo, kendo, aikido, or kyudo take a way to learn by looking at teacher without advice in details, so
called as “Mitori Keiko”. Like this, followers should figure out by themselves something important as tacit knowledge
from teacher. Although this may sound easy, unless an amateur identifies the elements of a master’s work that should
be observedand imitated, the skills will not be transmitted, as they rely on tacit knowledge.
The present research was made possible by the generosity of a master of Kazuwa-yaki,a form of pottery practised
in Kurayoshi City †, Tottori Prefecture, Japan. Kazuwa-yaki started in about 1750,and it disappeared for a time in the
1860s. In the 1930s, Otoyoshi Nakamori brought resurrected it in a form distinctive to Kyoto,after the potter’s wheel‡
and glaze§were introduced by second-generation artisans. A third-generation artisan, Hakuga Nakamori, now creates
innovative works while retaining the traditional and unique pinkish-red colour of Kazuwa-yaki pottery [6]. His work
received awards from the Japanese Kogei Association in 1997, 2000,and 2002.The present author filmed Mr.
Nakamori as he demonstrated how to use a potter’s wheel and then asked him to watch the recording. This article
analyses eye-movement data with the aim of identifying the tacit knowledge underlying the process by which amaster
potter passes on his techniques.
2. Previous studies
Cognitive science has highlighted the importance of vision in the process of recognition [7], and considerable
research has examined the role of vision in interpreting paintings and scenic representations. Some research has
focused on eye movements occurring as people try to solve aproblem or gather information when viewing an image
[8].
Human vision is characterised by several phenomena, including fixation, saccades (high-speed eye movements),
and blinking [9], which occur over about 100–400, 20–40, and 100–150 ms, respectively. Human beingsblink
approximately 10–15 times per minute. Due to saccadesand blinking, people are functionally blind for about 15% of
the time spent in visual activity [10]. These motions are known today as microsaccades (or fixational saccades), drifts,
and tremor [11]. Although visually acquired information is not always stable, at aminimum people are able to pay
conscious attention during fixation. Thus, this study was concerned with visual fixation.
Most research examining the effectiveness of advertising using eye tracking has considered print advertisements,
TV commercials, and banner advertisements on web pages [12,13,14]. In the marketing field, eye-tracking techniques
†Kurayoshi City is located in the western part of Japan. It is located 550 km away from Tokyo and is 6.5 km from the shore of the
Sea of Japan.
‡An apparatus that shapes pottery into a round form by spinning.
§ A preparation applied to pottery before it is heated in a kiln. Glaze prevents water leaks and lends a glossy appearance to the
pottery.

Jun Nakamura et al. / Procedia Computer Science 159 (2019) 1680–1687 1681
Available online at www.sciencedirect.com
ScienceDirect
Procedia Computer Science 00 (2019) 000–000
www.elsevier.com/locate/procedia
1877-0509 © 2019 The Author(s). Published by Elsevier B.V.
This is an open access article under the CC BY-NC-ND licen se (https://creativecommons.org/licenses/by-nc-nd/4.0/)
Peer review is the responsibility of KES International.
23rd International Conference on Knowledge-Based and Intelligent Information & Engineering
Systems
The pottery skills and tacit knowledge of a maser:
An analysis using eye-tracking data
Jun Nakamuraa*, Sanetake Nagayoshib
aChuo University, 742-1 Higashinakano, Hachio-ji, Tokyo 192-0393, Japan
bShizuoka University,
3-5-1 Johoku, Naka-ku, Hamamatsu, Shizuoka 432-8011
, Japan
Abstract
An important aspect of Japanese traditional culture is the design and creation of pots,bowls, and other pottery, an artform that has
historically been subject to cultural influences from China and the Korean peninsula. It has been suggested that passing down of
hand-crafting techniquesto the next generation is quite difficult, as the techniquesinvolve considerable tacit knowledge.
Thus, this study aimed to explore the creative processes of amaster of the craft using an eye-tracking system. We recorded the eye
movements of Hakuga Nakamori, a third-generation artisan, as he created a bowl and a pot.The results showed that, in the process
of working toward creating apiece of art, Mr.Nakamori first focused his attention on the whole piece,and then on certain elements
thereof.
© 2019 The Author(s). Published by Elsevier B.V.
This is an open access article under the CC BY-NC-ND license (https://creativecommons.org/licenses/by-nc-nd/4.0/)
Peer review is the responsibility of KES International.
Keywords: Eye tracking; cognitive process; moulding
1. Introduction
Japan has the longest history of pottery-making in the world, although the development of this artform has been
subject to cultural influences from China and the Korean peninsula [1]. The oldest pottery pieces in the world, which
are about 12,000 years old, were excavated in Japan. During the 4th to mid-5th centuries,the potter’s wheel, which
enabled artisans to create works of various shapes, was introduced to Japan from the Korean peninsula. Later, the kiln
* Corresponding author.
E-mail address: jyulis.77f@g.chuo-u.ac.jp, jyulis@tamacc.chuo-u.ac.jp
Available online at www.sciencedirect.com
ScienceDirect
Procedia Computer Science 00 (2019) 000–000
www.elsevier.com/locate/procedia
1877-0509 © 2019 The Author(s). Published by Elsevier B.V.
This is an open access article under the CC BY-NC-ND licen se (https://creativecommons.org/licenses/by-nc-nd/4.0/)
Peer review is the responsibility of KES International.
23rd International Conference on Knowledge-Based and Intelligent Information & Engineering
Systems
The pottery skills and tacit knowledge of a maser:
An analysis using eye-tracking data
Jun Nakamuraa*, Sanetake Nagayoshib
aChuo University, 742-1 Higashinakano, Hachio-ji, Tokyo 192-0393, Japan
bShizuoka University,
3-5-1 Johoku, Naka-ku, Hamamatsu, Shizuoka 432-8011
, Japan
Abstract
An important aspect of Japanese traditional culture is the design and creation of pots,bowls, and other pottery, an artform that has
historically been subject to cultural influences from China and the Korean peninsula. It has been suggested that passing down of
hand-crafting techniquesto the next generation is quite difficult, as the techniquesinvolve considerable tacit knowledge.
Thus, this study aimed to explore the creative processes of amaster of the craft using an eye-tracking system. We recorded the eye
movements of Hakuga Nakamori, a third-generation artisan, as he created a bowl and a pot.The results showed that, in the process
of working toward creating apiece of art, Mr.Nakamori first focused his attention on the whole piece,and then on certain elements
thereof.
© 2019 The Author(s). Published by Elsevier B.V.
This is an open access article under the CC BY-NC-ND license (https://creativecommons.org/licenses/by-nc-nd/4.0/)
Peer review is the responsibility of KES International.
Keywords: Eye tracking; cognitive process; moulding
1. Introduction
Japan has the longest history of pottery-making in the world, although the development of this artform has been
subject to cultural influences from China and the Korean peninsula [1]. The oldest pottery pieces in the world, which
are about 12,000 years old, were excavated in Japan. During the 4th to mid-5th centuries,the potter’s wheel, which
enabled artisans to create works of various shapes, was introduced to Japan from the Korean peninsula. Later, the kiln
* Corresponding author.
E-mail address: jyulis.77f@g.chuo-u.ac.jp, jyulis@tamacc.chuo-u.ac.jp
2Author name / Procedia Computer Science 00 (2019) 000–000
was introduced,providing a sophisticated system for heating pottery to 1,000°C. This solved the problem of water
leakage,such that the pottery showed durability to liquids.
As an expression of traditional Japanese culture,pottery making obviously requires old-fashioned hand-crafting
techniques,despite recent developments in digital technology.Some have suggested that it is very difficult to pass
such techniques on to the next generation [2,3]. This study aims to draw attention to this artform to avoid the loss of
pottery-making techniques that are based on tacit knowledge [4], and to preservethe traditional skillsfor future
generations. Verbalising the techniques by which these skills could be conveyed is difficult [5], so this study focused
on eye movements. When recognising objects, people unconsciously come to an understanding of the ‘what’ of the
object; asthe proverb goes, “The eyes are the windows to the soul”. Another old proverb suggests that one should
“learn by seeing his back”, i.e. that learners should acquire skills by observing veterans. For an example, Japanese
martial arts like judo, kendo, aikido, or kyudo take a way to learn by looking at teacher without advice in details, so
called as “Mitori Keiko”. Like this, followers should figure out by themselves something important as tacit knowledge
from teacher. Although this may sound easy, unless an amateur identifies the elements of a master’s work that should
be observedand imitated, the skills will not be transmitted, as they rely on tacit knowledge.
The present research was made possible by the generosity of a master of Kazuwa-yaki,a form of pottery practised
in Kurayoshi City †, Tottori Prefecture, Japan. Kazuwa-yaki started in about 1750,and it disappeared for a time in the
1860s. In the 1930s, Otoyoshi Nakamori brought resurrected it in a form distinctive to Kyoto,after the potter’s wheel‡
and glaze§were introduced by second-generation artisans. A third-generation artisan, Hakuga Nakamori, now creates
innovative works while retaining the traditional and unique pinkish-red colour of Kazuwa-yaki pottery [6]. His work
received awards from the Japanese Kogei Association in 1997, 2000,and 2002.The present author filmed Mr.
Nakamori as he demonstrated how to use a potter’s wheel and then asked him to watch the recording. This article
analyses eye-movement data with the aim of identifying the tacit knowledge underlying the process by which amaster
potter passes on his techniques.
2. Previous studies
Cognitive science has highlighted the importance of vision in the process of recognition [7], and considerable
research has examined the role of vision in interpreting paintings and scenic representations. Some research has
focused on eye movements occurring as people try to solve aproblem or gather information when viewing an image
[8].
Human vision is characterised by several phenomena, including fixation, saccades (high-speed eye movements),
and blinking [9], which occur over about 100–400, 20–40, and 100–150 ms, respectively. Human beingsblink
approximately 10–15 times per minute. Due to saccadesand blinking, people are functionally blind for about 15% of
the time spent in visual activity [10]. These motions are known today as microsaccades (or fixational saccades), drifts,
and tremor [11]. Although visually acquired information is not always stable, at aminimum people are able to pay
conscious attention during fixation. Thus, this study was concerned with visual fixation.
Most research examining the effectiveness of advertising using eye tracking has considered print advertisements,
TV commercials, and banner advertisements on web pages [12,13,14]. In the marketing field, eye-tracking techniques
†Kurayoshi City is located in the western part of Japan. It is located 550 km away from Tokyo and is 6.5 km from the shore of the
Sea of Japan.
‡An apparatus that shapes pottery into a round form by spinning.
§ A preparation applied to pottery before it is heated in a kiln. Glaze prevents water leaks and lends a glossy appearance to the
pottery.

1682 Jun Nakamura et al. / Procedia Computer Science 159 (2019) 1680–1687
Author name / Procedia Computer Science 00 (2019) 000–000 3
have been applied in numerous ways, and one study focused on the eye movements of individuals tasked with
shopping within alimited amount of time [15].
However, few studies on the transfer of traditional artistic techniques have employed eye tracking.As an example
[16], a special interest group on knowledge, skill,and technology (SIG-KST) barely discussed such techniques. Indeed,
after 36 conferences between the group’s founding in June 2007 and March 2019, only one associated publication
employed eye tracking [17]. Thus,there is gap with respect to examination of the transfer of traditional techniquesvia
an eye tracking paradigm.
Taguchi (2013) summarised methods for passing down techniques among Japanese domestic corporations, and
divided them into eight types:1) technical documents;2) video libraries; 3) Curriculum Development Based on Ability
Structure (CUDBAS); 4) on-the-job training (OJT); 5) off-the-job training (Off-JT); 6. “self-job training”(SJT); 7.
IT-based systems; and 8) consultation [18]. The present paper is most closely related to video libraries,for which a
weak point is that the value of avideo depends on the knowledge and ability of the viewers. This paper aimed to
address that weak point.
3. The process of making pottery
Completing a pottery piece requires a series of procedures. The first step is to prepare the basic material, i.e. the
clay, and the next step is to mould and shape it.Of the several techniques available for shaping clay, this paper focuses
on shaping a piece using a potter’s wheel. The third step is to separate the pottery from the wheel, add patterns to the
surface, and dry it to eliminate any moisture. After it is heated in the kiln at 700–800°C, the piece is painted with glaze.
Subsequently, it is heated again, this time at 1,100–1,300°C, causing the surface to turn red due to the cinnabar in the
glaze. Finally,it is baked again at low temperatures and then removed from the kiln.
This paper concentrates on the moulding step,in recognition of Mr. Nakamori’s suggestion that this is one of the
processes that differs most markedly among artisans, and also and that it is difficult to articulate,being based on the
tacit knowledge of the master. It was hypothesised that the tacit knowledge would be expressed as follows:
1) By the potter first concentrating on the entirety of the clay on the potter’s wheel ; and then
2) Concentrating on a particular part of the spinning lump of clay; where
3) Steps 1) and 2) would then be repeated throughout the moulding process.
The hypotheses above are based on previous research [16] on system operationsthat identified the three steps of
first paying attention to the whole, then focusing on a particular element, and finally concentrating on relatively large
partsof the whole.
4. Study design
4.1. Preparation
Prior to explaining the methods employed herein, I will first introduce the equipment that was used for the study,
which included a video camera for filming, an eye-tracking system, a monitor, a computer, and software to film and
record eye movements:
-Video camera:HDR-AS300 (Sony)
-Eye tracker: AB:X2-30 (Tobii Technology)
-Monitor:23.8-in anomaly detection system,colour display (Princeton)
-Software: in-house developed system architecture (Figure 1), with user interface (Figure 2)
-The following steps were carried out:
Adatabase was made to store the video files
xand yeye-tracking coordinate data were recorded for every video
4Author name / Procedia Computer Science 00 (2019) 000–000
Eye tracking data was shown on the video along with timeline
Unit of coordinate on the video was not a pixel but grill, which was chosen in the analysis setting.
The xand ycoordinates for both eyes was shown for 1000 mm per second from the beginning of eye tracking.
Fig. 1. Developed system architecture. Fig. 2. User interface.
4.2. Overview of the study
Visual target:
Roughly shaped pot or bowl
Duration:
Approximately 5 minutes
Artisan:
Hakuga Nakamori
Location:
Kurashiki City, Tottori Prefecture, Japan
Date:
Tuesday, 5 February 2019
Time:
9:30–11:30 (Japan Standard Time)
Flow: 1) Film the process of shaping the pottery (pot followed by abowl)
2) Play the video on the screen and record Mr. Nakamori’s eye movements
3) Analyse the eye-tracking data
Mr. Nakamori’s eye movements were measured as he watched the video.The eye-tracking glasses necessary to
capture eye movements during the actual process of creating the pottery were not available.As an alternative, asking
Mr. Nakamori to wear an action camera was considered, but ultimately this was not done because it would be unnatural
to sculpt pottery with the head in a completely stable and frozen position.Thus, this study should be regarded as
preliminary, early stage research.
In analysing the eye-tracking data according to the hypotheses delineated above, a level of attention was calculated
in the similar manner [19,20].
Based on the history of the xand ycoordinates at time ‘t’, the diversity in attentional focus is calculated as follows:
 =(() ())+(() ()) (1)
 =(() ( ))+(() ())

 (2)
{()}=/ (1) ( /) (3)
,where
-(() ())is a two dimensional vector that indicates the coordinate location on the screen at a time of 
- is the straight-line distance between the start at a time of and the end of the tracking point at a time of 
- is the accumulated distance covered by all eye movement

Jun Nakamura et al. / Procedia Computer Science 159 (2019) 1680–1687 1683
Author name / Procedia Computer Science 00 (2019) 000–000 3
have been applied in numerous ways, and one study focused on the eye movements of individuals tasked with
shopping within alimited amount of time [15].
However, few studies on the transfer of traditional artistic techniques have employed eye tracking.As an example
[16], a special interest group on knowledge, skill,and technology (SIG-KST) barely discussed such techniques. Indeed,
after 36 conferences between the group’s founding in June 2007 and March 2019, only one associated publication
employed eye tracking [17]. Thus,there is gap with respect to examination of the transfer of traditional techniquesvia
an eye tracking paradigm.
Taguchi (2013) summarised methods for passing down techniques among Japanese domestic corporations, and
divided them into eight types:1) technical documents;2) video libraries; 3) Curriculum Development Based on Ability
Structure (CUDBAS); 4) on-the-job training (OJT); 5) off-the-job training (Off-JT); 6. “self-job training”(SJT); 7.
IT-based systems; and 8) consultation [18]. The present paper is most closely related to video libraries,for which a
weak point is that the value of avideo depends on the knowledge and ability of the viewers. This paper aimed to
address that weak point.
3. The process of making pottery
Completing a pottery piece requires a series of procedures. The first step is to prepare the basic material, i.e. the
clay, and the next step is to mould and shape it.Of the several techniques available for shaping clay, this paper focuses
on shaping a piece using a potter’s wheel. The third step is to separate the pottery from the wheel, add patterns to the
surface, and dry it to eliminate any moisture. After it is heated in the kiln at 700–800°C, the piece is painted with glaze.
Subsequently, it is heated again, this time at 1,100–1,300°C, causing the surface to turn red due to the cinnabar in the
glaze. Finally,it is baked again at low temperatures and then removed from the kiln.
This paper concentrates on the moulding step,in recognition of Mr. Nakamori’s suggestion that this is one of the
processes that differs most markedly among artisans, and also and that it is difficult to articulate,being based on the
tacit knowledge of the master. It was hypothesised that the tacit knowledge would be expressed as follows:
1) By the potter first concentrating on the entirety of the clay on the potter’s wheel ; and then
2) Concentrating on a particular part of the spinning lump of clay; where
3) Steps 1) and 2) would then be repeated throughout the moulding process.
The hypotheses above are based on previous research [16] on system operationsthat identified the three steps of
first paying attention to the whole, then focusing on a particular element, and finally concentrating on relatively large
partsof the whole.
4. Study design
4.1. Preparation
Prior to explaining the methods employed herein, I will first introduce the equipment that was used for the study,
which included a video camera for filming, an eye-tracking system, a monitor, a computer, and software to film and
record eye movements:
-Video camera:HDR-AS300 (Sony)
-Eye tracker: AB:X2-30 (Tobii Technology)
-Monitor:23.8-in anomaly detection system,colour display (Princeton)
-Software: in-house developed system architecture (Figure 1), with user interface (Figure 2)
-The following steps were carried out:
Adatabase was made to store the video files
xand yeye-tracking coordinate data were recorded for every video
4Author name / Procedia Computer Science 00 (2019) 000–000
Eye tracking data was shown on the video along with timeline
Unit of coordinate on the video was not a pixel but grill, which was chosen in the analysis setting.
The xand ycoordinates for both eyes was shown for 1000 mm per second from the beginning of eye tracking.
Fig. 1. Developed system architecture. Fig. 2. User interface.
4.2. Overview of the study
Visual target:
Roughly shaped pot or bowl
Duration:
Approximately 5 minutes
Artisan:
Hakuga Nakamori
Location:
Kurashiki City, Tottori Prefecture, Japan
Date:
Tuesday, 5 February 2019
Time:
9:30–11:30 (Japan Standard Time)
Flow: 1) Film the process of shaping the pottery (pot followed by abowl)
2) Play the video on the screen and record Mr. Nakamori’s eye movements
3) Analyse the eye-tracking data
Mr. Nakamori’s eye movements were measured as he watched the video.The eye-tracking glasses necessary to
capture eye movements during the actual process of creating the pottery were not available.As an alternative, asking
Mr. Nakamori to wear an action camera was considered, but ultimately this was not done because it would be unnatural
to sculpt pottery with the head in a completely stable and frozen position.Thus, this study should be regarded as
preliminary, early stage research.
In analysing the eye-tracking data according to the hypotheses delineated above, a level of attention was calculated
in the similar manner [19,20].
Based on the history of the xand ycoordinates at time ‘t’, the diversity in attentional focus is calculated as follows:
 =(() ())+(() ()) (1)
 =(() ( ))+(() ())

 (2)
{()}=/ (1) ( /) (3)
,where
-(() ())is a two dimensional vector that indicates the coordinate location on the screen at a time of 
- is the straight-line distance between the start at a time of and the end of the tracking point at a time of 
- is the accumulated distance covered by all eye movement

1684 Jun Nakamura et al. / Procedia Computer Science 159 (2019) 1680–1687
Author name / Procedia Computer Science 00 (2019) 000–000 5
The saccade is a movement defined with more than 2.0pxl/ms in this paper. In addition, within the saccade, the
movement with less than 10deg/sec is called as saccade-line here. The number of focus points and the range of twere
both adjustable in this system.
5. Results
Figure 3 displays one result of the experiment, showing the process of moulding the upper part of abowl.The x
axis is the amount of time and y axis is the result of calculation of Eq. (3).
Fig. 3. Resultsof Eq.(3). Upper panel:time spent focusing on the bowl; lower panel: time spent focusing on the base.
According to the graph, the eye movements initially corresponded to the attainment of an overview,and then the
visual attention was narrowed to focus on aparticular part of the piece; this was clear despite some missing data
(between 25 and 36 sfor the pot and 31 and 37 sfor the bowl).
The bowl-moulding process can be characterised as follows:
Total time: 427.211 ms (7.12 minutes)
Number of focus points: 1,420
Grid size:x:18, y:18
Maximum E score,left eye: 0.121420 (SD: 0.014071)
Maximum Escore,right eye: 0.122676 (SD: 0.017431)
As Figure 4 shows, saccade-line were observed about 42 seconds after the start of the moulding process; at
approximately the same time,the eyes captured the whole piece. Then, at 1 minute 12 seconds, the eyesmoved to the
left thumb and hand,as shown in Figure 5.
The pot-moulding process can be characterised as follows:
Total time: 477.617ms (7.96 minutes)
6Author name / Procedia Computer Science 00 (2019) 000–000
Number of focus points: 1,439
Grid size: x:18, y:18
Maximum Escore,left eye: 0.122756 (SD: 0.012097)
Maximum Escore,right eye: 0.121667 (SD: 0.014829)
Fig. 4. Observed saccade-line (bowl). Fig. 5. Eyes focus on the hands (bowl).
Mr Nakamori’s eyes fixated on patterned areas at 47 seconds after the start of moulding process,at which time his
fingers were moving continuously from the centre outward (Figure 6). At 1 minute 4 seconds,his vision was directed
to the tips of the fingers on the left hand (Figure 7).
Fig. 6. Shit in eye fixation point (pot). Fig. 7. Eyes focus on the left hand (pot).

Jun Nakamura et al. / Procedia Computer Science 159 (2019) 1680–1687 1685
Author name / Procedia Computer Science 00 (2019) 000–000 5
The saccade is a movement defined with more than 2.0pxl/ms in this paper. In addition, within the saccade, the
movement with less than 10deg/sec is called as saccade-line here. The number of focus points and the range of twere
both adjustable in this system.
5. Results
Figure 3 displays one result of the experiment, showing the process of moulding the upper part of abowl.The x
axis is the amount of time and y axis is the result of calculation of Eq. (3).
Fig. 3. Resultsof Eq.(3). Upper panel:time spent focusing on the bowl; lower panel: time spent focusing on the base.
According to the graph, the eye movements initially corresponded to the attainment of an overview,and then the
visual attention was narrowed to focus on aparticular part of the piece; this was clear despite some missing data
(between 25 and 36 sfor the pot and 31 and 37 sfor the bowl).
The bowl-moulding process can be characterised as follows:
Total time: 427.211 ms (7.12 minutes)
Number of focus points: 1,420
Grid size:x:18, y:18
Maximum E score,left eye: 0.121420 (SD: 0.014071)
Maximum Escore,right eye: 0.122676 (SD: 0.017431)
As Figure 4 shows, saccade-line were observed about 42 seconds after the start of the moulding process; at
approximately the same time,the eyes captured the whole piece. Then, at 1 minute 12 seconds, the eyesmoved to the
left thumb and hand,as shown in Figure 5.
The pot-moulding process can be characterised as follows:
Total time: 477.617ms (7.96 minutes)
6Author name / Procedia Computer Science 00 (2019) 000–000
Number of focus points: 1,439
Grid size: x:18, y:18
Maximum Escore,left eye: 0.122756 (SD: 0.012097)
Maximum Escore,right eye: 0.121667 (SD: 0.014829)
Fig. 4. Observed saccade-line (bowl). Fig. 5. Eyes focus on the hands (bowl).
Mr Nakamori’s eyes fixated on patterned areas at 47 seconds after the start of moulding process,at which time his
fingers were moving continuously from the centre outward (Figure 6). At 1 minute 4 seconds,his vision was directed
to the tips of the fingers on the left hand (Figure 7).
Fig. 6. Shit in eye fixation point (pot). Fig. 7. Eyes focus on the left hand (pot).

1686 Jun Nakamura et al. / Procedia Computer Science 159 (2019) 1680–1687
Author name / Procedia Computer Science 00 (2019) 000–000 7
6. Discussion
The previous sections described some characteristic elements of the moulding process.First,Mr Nakamori’s
fixationsallowed a wide view to be captured,including both the clay and the potter’s wheel; subsequently,his eyes
focused on the left hand and then the tips of the fingers. This sequence was very similar to that laid out above in the
three study hypotheses.
Second, in the process of shifting his concentration from the whole to a part of the piece,Mr Nakamori’s eyes
tracked from the centre outward when moulding both the pot and the bowl,showing that both the shape of the clay
and the position of the fingers were checked during themoulding process.Although the moulding of a pot and a bowl
share common features, some differences in the data were also seen;for instance, an obvious saccade-line occurred
while moulding the pot (Figure 4), and Mr Nakamori’s eyesshowed aleftward movement pattern during the bowl-
moulding process (see Figure 6). Authors have interviewed Mr.Nakamori on this habit that pays attention to the left
hand side. He explained that the edge is so important when it comes to make sure of shaping outer line.
Fig. 8. Moulding with a stick (pot)
Finally,Mr Nakamori concentrated on the tips of the fingers,which appears to be of key importance with respect
to revealing his tacit knowledge. Some elements of the moulding process are impossible to capture with the eye
tracker;for example, eye tracking cannot reveal when and to what degree the fingers apply pressure. When creating a
bowl, a stick is used to apply pressure from inside while the left hand, placed on the outer surface,senses the thickness
of the bowl;the potter’s tacit knowledge guides further shaping efforts (Figure 8).
Thus, acomplete formalisation of the tacit knowledge associated with the process of creating pottery remains a
distant goal, but the eye-tracking system used in the present study nevertheless extracted some key elements of this
process.
7. Conclusion
Using eye-tracking technology, this study examined an artisan’s techniques for crafting traditional Japanese pottery,
which usually remain as undefined elements within a body of tacit knowledge. A key component was the shift of focus
to the tips of the fingers;it is important that techniques for controlling the finger tips be verbalised.Although many
8Author name / Procedia Computer Science 00 (2019) 000–000
studies have used eye tracking in the context of static images, dynamic eye-movement data coupled with tacit
knowledge would facilitate the process of passing on culturally important traditional techniques to the next generation.
Thecreation of a general instructional manual would be one way to convey such information, but demonstrating when
the shiftsin focus occur based on eye-tracking data would clearly be helpful.I intend to collect additional eye-tracking
data towards this end.
A video-based risk-recognition training system [21]revealed differences in eye movements between veterans and
amateurs. However, the focus of that study was on risk prediction rather than the transmission of techniques. In future
work, also intend to reveal differences between the eye movements of a master and an amateur,albeit with a different
aim in mind.
Acknowledgements
We are very grateful to Mr. Hakuga Nakamori, a third-generation master of Kazuwa-yaki, for his invaluable
cooperation with this research. We also want to thank Mr. Kenta Otsuka, a student in my seminar group who is
currently working for NTT-Data and who assisted us in preparing and conducting this study. Last but not least, we
thank Mr. Naomasa Okumura and Ms. Reiko Fujiwara for introducing us to Mr. Hakuga Nakamori. In addition, this
research was supported by the grant from JSPS (JSPS KAKENHI Grant Number 19K03062).
References
[1] http://www.liqun123.com/ (viewed on 30 March 2019)
[2] Furukawa, K. (2012) “Gijutu Ginou Denshou Mattanashi (No enough time for technology transfer).” Factory Management 58 (14):24-27.
[3] Taruta, Y. (2012) “Consideration for succession of knowledge, technology, and skill.” SIG-KST 2012-034-04: 1-6.
[4] Polan yi, M.,and Sen, A. (2009) “The Tacit Dimension.” University of Chicago Press, Reissue edition
[5] Johansson, P., Hall, L., Sikstrom, S., and Olsson.(2005) “A. Failure to detect mismatches between intention and outcome in a simple decision
task.” Science 310:116-119, 2005.
[6] https://www.pref.tottori.lg.jp/30635.htm (viewed on 30 March 2019)
[7] Salvucci, DD., and Anderson, JR. (1998) “Tracing eye movement protocols with cognitive process models.” Proc. Annual Conference of the
Cognitive Science Society:923-928.
[8] Terai, H., and Miwa, L. (2003) “Insight problem solving from the viewpoint of constraint relaxation using eye movement analysis.” Proc. of
International Conference of Cognitive Science:671-676.
[9] Martinez-Conde, S. (2004) “The role of fixational eye movements in visual perception.” Nature Previews Neuroscience 5:229-240.
[10] Van Der Lans, R., and Wedel, M. (2017) “Eye movements during search and choice.” Handbook of Marketing Decision Models, 2nd edition,
Wierenga, B., and Van der Lans, R. (eds.), Springer:331-359.
[11] Martinez-Conde, and S., Macknik, S.L. (2015) “From Exploration to Fixation: An Integrative View of Yarbus’s Vision.” Perception 44 (8-
9):884-899.
[12] Wedel, M., and Pieters, R. (2006) “Eye tracking for visual marketing.” Foundations and Trends in Marketing 1:231–320.
[13] Wedel, M., and Pieters, R. (2008) “A review of eye-tracking research in marketing.” Review of Marketing Research, 4, K. N. Malhotra (ed.),
Emerald Group Publishing:122–147.
[14] Dr`eze, X., and Hussherr, F.X. (2003) “Internet advertising: Is anybody watching?”Journal of Interactive Marketing 17: 8–23.
[15] Pieters, R., Warlop, L. (1999) “Visual attention during brand choice: the impact of time pressure and task motivation.” International Journal
of Research in Marketing 16 (1): 1-16.
[16] Furukawa, Y. (2014) “Tishiki gijutu ginou no denshou ni kansuru kousatu (Discussion about transfer support of knowledge, skill and
technology).”, SIG-KST 2014-01-04.
[17] Shirayama, S., Takato, K., Osawa, R., and Bonkohara, N. (2017) “Abasic study of quantitative description of how to fly a taketombo.” SIG-
KST-030-06.
[18] Taguchi, Y.(2013) “Rokunai kigyo no ginoudensho no torikumi ni kannsuru itikosatu (Approaches to pass down techniques among
Japanese domestic corporations).” Bulletin of Shohoku College
[19] Nakamura, J., Ohsawa, Y. (2008) “Uncertainty items as concept sourcesin analogical thinking.” Proc.of International Conference on
Cognitive Science,Moscow:111-113.
[20]Krejtz, K., Duchowski, A., Szmidt, T., Krejtz, I., González Perilli, F., Pires, A., Vilaro, A., and Villalobos, N. (2015) “Gaze transition
entropy.” ACM Transactions on Applied Perception 13 (1) 4:1-4:20.
[21] Fujita, Y., Nakamura, J., and Kushiro, N. (2017) “Risk recognition training system by movie—characteristics of risk recognition from
experts’ gaze.” Proc. of Forum on Information Technology

Jun Nakamura et al. / Procedia Computer Science 159 (2019) 1680–1687 1687
Author name / Procedia Computer Science 00 (2019) 000–000 7
6. Discussion
The previous sections described some characteristic elements of the moulding process.First,Mr Nakamori’s
fixationsallowed a wide view to be captured,including both the clay and the potter’s wheel; subsequently,his eyes
focused on the left hand and then the tips of the fingers. This sequence was very similar to that laid out above in the
three study hypotheses.
Second, in the process of shifting his concentration from the whole to a part of the piece,Mr Nakamori’s eyes
tracked from the centre outward when moulding both the pot and the bowl,showing that both the shape of the clay
and the position of the fingers were checked during themoulding process.Although the moulding of a pot and a bowl
share common features, some differences in the data were also seen;for instance, an obvious saccade-line occurred
while moulding the pot (Figure 4), and Mr Nakamori’s eyesshowed aleftward movement pattern during the bowl-
moulding process (see Figure 6). Authors have interviewed Mr.Nakamori on this habit that pays attention to the left
hand side. He explained that the edge is so important when it comes to make sure of shaping outer line.
Fig. 8. Moulding with a stick (pot)
Finally,Mr Nakamori concentrated on the tips of the fingers,which appears to be of key importance with respect
to revealing his tacit knowledge. Some elements of the moulding process are impossible to capture with the eye
tracker;for example, eye tracking cannot reveal when and to what degree the fingers apply pressure. When creating a
bowl, a stick is used to apply pressure from inside while the left hand, placed on the outer surface,senses the thickness
of the bowl;the potter’s tacit knowledge guides further shaping efforts (Figure 8).
Thus, acomplete formalisation of the tacit knowledge associated with the process of creating pottery remains a
distant goal, but the eye-tracking system used in the present study nevertheless extracted some key elements of this
process.
7. Conclusion
Using eye-tracking technology, this study examined an artisan’s techniques for crafting traditional Japanese pottery,
which usually remain as undefined elements within a body of tacit knowledge. A key component was the shift of focus
to the tips of the fingers;it is important that techniques for controlling the finger tips be verbalised.Although many
8Author name / Procedia Computer Science 00 (2019) 000–000
studies have used eye tracking in the context of static images, dynamic eye-movement data coupled with tacit
knowledge would facilitate the process of passing on culturally important traditional techniques to the next generation.
Thecreation of a general instructional manual would be one way to convey such information, but demonstrating when
the shiftsin focus occur based on eye-tracking data would clearly be helpful.I intend to collect additional eye-tracking
data towards this end.
A video-based risk-recognition training system [21]revealed differences in eye movements between veterans and
amateurs. However, the focus of that study was on risk prediction rather than the transmission of techniques. In future
work, also intend to reveal differences between the eye movements of a master and an amateur,albeit with a different
aim in mind.
Acknowledgements
We are very grateful to Mr. Hakuga Nakamori, a third-generation master of Kazuwa-yaki, for his invaluable
cooperation with this research. We also want to thank Mr. Kenta Otsuka, a student in my seminar group who is
currently working for NTT-Data and who assisted us in preparing and conducting this study. Last but not least, we
thank Mr. Naomasa Okumura and Ms. Reiko Fujiwara for introducing us to Mr. Hakuga Nakamori. In addition, this
research was supported by the grant from JSPS (JSPS KAKENHI Grant Number 19K03062).
References
[1] http://www.liqun123.com/ (viewed on 30 March 2019)
[2] Furukawa, K. (2012) “Gijutu Ginou Denshou Mattanashi (No enough time for technology transfer).” Factory Management 58 (14):24-27.
[3] Taruta, Y. (2012) “Consideration for succession of knowledge, technology, and skill.” SIG-KST 2012-034-04: 1-6.
[4] Polan yi, M.,and Sen, A. (2009) “The Tacit Dimension.” University of Chicago Press, Reissue edition
[5] Johansson, P., Hall, L., Sikstrom, S., and Olsson.(2005) “A. Failure to detect mismatches between intention and outcome in a simple decision
task.” Science 310:116-119, 2005.
[6] https://www.pref.tottori.lg.jp/30635.htm (viewed on 30 March 2019)
[7] Salvucci, DD., and Anderson, JR. (1998) “Tracing eye movement protocols with cognitive process models.” Proc. Annual Conference of the
Cognitive Science Society:923-928.
[8] Terai, H., and Miwa, L. (2003) “Insight problem solving from the viewpoint of constraint relaxation using eye movement analysis.” Proc. of
International Conference of Cognitive Science:671-676.
[9] Martinez-Conde, S. (2004) “The role of fixational eye movements in visual perception.” Nature Previews Neuroscience 5:229-240.
[10] Van Der Lans, R., and Wedel, M. (2017) “Eye movements during search and choice.” Handbook of Marketing Decision Models, 2nd edition,
Wierenga, B., and Van der Lans, R. (eds.), Springer:331-359.
[11] Martinez-Conde, and S., Macknik, S.L. (2015) “From Exploration to Fixation: An Integrative View of Yarbus’s Vision.” Perception 44 (8-
9):884-899.
[12] Wedel, M., and Pieters, R. (2006) “Eye tracking for visual marketing.” Foundations and Trends in Marketing 1:231–320.
[13] Wedel, M., and Pieters, R. (2008) “A review of eye-tracking research in marketing.” Review of Marketing Research, 4, K. N. Malhotra (ed.),
Emerald Group Publishing:122–147.
[14] Dr`eze, X., and Hussherr, F.X. (2003) “Internet advertising: Is anybody watching?”Journal of Interactive Marketing 17: 8–23.
[15] Pieters, R., Warlop, L. (1999) “Visual attention during brand choice: the impact of time pressure and task motivation.” International Journal
of Research in Marketing 16 (1): 1-16.
[16] Furukawa, Y. (2014) “Tishiki gijutu ginou no denshou ni kansuru kousatu (Discussion about transfer support of knowledge, skill and
technology).”, SIG-KST 2014-01-04.
[17] Shirayama, S., Takato, K., Osawa, R., and Bonkohara, N. (2017) “Abasic study of quantitative description of how to fly a taketombo.” SIG-
KST-030-06.
[18] Taguchi, Y.(2013) “Rokunai kigyo no ginoudensho no torikumi ni kannsuru itikosatu (Approaches to pass down techniques among
Japanese domestic corporations).” Bulletin of Shohoku College
[19] Nakamura, J., Ohsawa, Y. (2008) “Uncertainty items as concept sourcesin analogical thinking.” Proc.of International Conference on
Cognitive Science,Moscow:111-113.
[20]Krejtz, K., Duchowski, A., Szmidt, T., Krejtz, I., González Perilli, F., Pires, A., Vilaro, A., and Villalobos, N. (2015) “Gaze transition
entropy.” ACM Transactions on Applied Perception 13 (1) 4:1-4:20.
[21] Fujita, Y., Nakamura, J., and Kushiro, N. (2017) “Risk recognition training system by movie—characteristics of risk recognition from
experts’ gaze.” Proc. of Forum on Information Technology