Boccia Court Analisys for Real-time Scoring

Pedro Leite, Alexandre Calado and Filomena Soares

Algoritmi Centre, University of Minho, Guimarães, Portugal

Keywords: Boccia, Object Tracking, Sports Analysis, Physical Inactivity, Computer Vision.

Abstract: This paper aims to develop a tool to increase the engagement and commitment of the elderly population in

the Boccia game. This tool is based on the implementation of a real-time computer vision algorithm that

analyses the Boccia court field and displays the score in a graphical user interface. In Portugal, physical

inactivity is associated with 14% of deaths per year, higher percentage than the world average, which is less

than 10%. In this paper, the Boccia game is used for promoting physical activity in the elderly, due to its

simplicity and adaptability to their physical impairments. This system intends to encourage the elders to lead

a more active lifestyle, providing a healthier life and possibly reducing the risk of major diseases and injuries.

1 INTRODUCTION

Physical inactivity is known to be the fourth leading

risk factor for death worldwide (WHO, 2010).

Regarding disease burdens, according to the World

Health Organization (WHO, 2015), 5% from

coronary heart disease, 7% from type 2 diabetes, 9%

from breast cancer and 10% colon cancer are

estimated to be the consequence of a deficit in

physical activity practice, thus resulting in 1 million

deaths (10% of deaths) only in the WHO European

Region.

In Portugal, it is estimated that physical inactivity

is associated with 14% of deaths per year, which is a

higher percentage than the world average (lower than

10%) (Direção-Geral da Saúde, 2017). Besides, it is

estimated that in a country with 10 million

inhabitants, which is approximately Portugal’s

population, if 50% are insufficiently active, there is a

cost associated with physical inactivity of €910

million per year (WHO, 2015), which is

approximately 9% of the Portuguese Ministry of

Health’s budget in 2017 (Direção-Geral da Saúde,

2017).

Amongst all age groups, the elderly is one of the

most physically inactive. (Matthews et al., 2008) have

identified adults older than 60 years to be the most

sedentary group in the United States and, moreover,

it is suggested that 50% of sedentary adults have no

intention of starting an exercise plan (Schutzer and

Graves, 2004). According to (Instituto Nacional de

Estatística, 2016), this reality is not different in

Portugal. In a country where the majority of the

population aged more than 15 years does not practice

any form of physical activity on a regular basis (5,8

million), only 19% of the population aged more than

65 years practices physical activity regularly.

Facing these concerning statistics, along with the

current increase in older population, comes the need

of developing new strategies for motivating the

elderly to engage more frequently in physical activity,

which can be very beneficial at such age. Amongst

the panoply of benefits that physical activity can

offer, the prevention of functional loss (Stessman et

al., 2009), reduction of the risk of falling (Gillespie et

al., 2012), blood pressure control (Westhoff et al.,

2007), improvement of the bones and joints’ health

(Lee et al., 2012) and maintenance of mental health

(Salguero et al., 2011) are particularly important for

the elderly, along with lower risks for all-cause

mortality (Landi et al., 2004).

The (WHO, 2010) recommends adults aged 65

years and above to engage in moderate-intensity

aerobic physical activity for at least 150 minutes or

engage in vigorous-intensity aerobic physical activity

for at least 75 minutes throughout the week.

Alternatively, an equivalent combination of moderate

and vigorous intensity activity can be done. However,

when such amounts of physical activity cannot be

reached due to health conditions, which happens

often at a late age, the individual should be as

physically active as his/her limitations and abilities

allow. As mentioned in (Landi et al., 2004), even

simple everyday activities, such as walking,

Leite, P., Calado, A. and Soares, F.

Boccia Court Analisys for Real-time Scoring.

DOI: 10.5220/0006918305110516

In Proceedings of the 15th International Conference on Informatics in Control, Automation and Robotics (ICINCO 2018) - Volume 2, pages 511-516

ISBN: 978-989-758-321-6

511

gardening or housekeeping seem to offer

considerable benefits for frail and older individuals.

Bearing this in mind, there are several examples

of sports and games that are not physically

demanding and can be played by the elderly with

reduced mobility, such as darts, pétanque or even

bowling (using a plastic ball). Another excellent

example is the Boccia game, a simple precision ball

sport similar to pétanque that was initially designed

to be played by individuals with cerebral palsy. Due

to its easy adaptability to the physical limitations of

the players, later it became a game played by persons

with other disabilities affecting motor skills.

Furthermore, as a team sport, it promotes social

interaction and motivates individuals to participate

more often (Estabrooks and Carron, 1999). All these

characteristics make Boccia an excellent game to

encourage physical activity practice amongst the

elderly.

Based on the aforementioned remarks, this paper

proposes a system that uses a computer vision

algorithm based on object tracking to automatically

compute and display the game score of a Boccia

match in real-time. The main objective of this system

is to make the experience of playing Boccia more fun

and enjoyable while promoting its practice, therefore

motivating the elderly to engage in physical activity.

In a later stage, the proposed system will be tested

during a Boccia match played in an actual nursing

home, thus its development was based on the

modified game rules used by the caregivers, along

with the characteristics of the room where the game

is usually played.

In future work, the proposed system will also be

integrated in the iBoccia framework (Figueira et al.,

2017; Silva et al., 2018) which features the use of

inertial sensors along with Microsoft Kinect to

monitor the player’s physical activity throughout a

game of Boccia. The extracted data will be stored in

a database and processed afterwards to help enhance

the player’s performance in the game and provide

relevant information for the caregiver such as the

detection of physical or cognitive decline.

This paper is organized as follows: section 2

offers a state-of-the-art regarding object tracking in

sports. It also presents a brief description of the

Boccia game and its rules, along with the modified

rules that are used in the nursing home. Section 3

presents a thorough description of the system’s

architecture. Section 4 describes the preliminary

results from the tests performed using the proposed

system. Finally, section 5 presents the conclusions

taken from the obtained results, along with future

work.

2 BACKGROUND

The Boccia game has been used as the context for a

limited number of studies regarding motion analysis

of the throwing movement involving individuals with

cerebral palsy (Huang et al., 2014; Tsai et al., 2014),

as well as individuals suffering from other disabilities

affecting motor skills (Arroxellas et al., 2017).

Regarding Boccia scoring systems, (Wang et al.,

2017) developed a computer competition system and

a computer scoreboard system to be used in official

Boccia events and competitions. However, within the

Boccia context, no literature was found concerning

object tracking or real-time game score computation.

Nevertheless, in the last years there has been a

growing trend for using computer vision in sports

analysis.

Object tracking is particularly relevant for sports

analysis and different tracking techniques have

already been used in various works. For instance,

(Pingali et al., 2000) implemented multi-camera

estimation for tracking the motion of a tennis ball in

3D. In this technique, six fixed cameras are placed

around the area of play and the resulting images are

processed afterwards to mount a 3D image which

allows the tracking of the ball’s entire trajectory. This

technique is also used by the Hawk-Eye system

(“Hawk-Eye”, 2018), which is currently applied in a

wide variety of sports, such as football, cricket and

tennis.

(Wu et al., 2006) developed an algorithm for the

detection of a basketball based on its colour and

shape, using clustering. If the detection was

succeeded, then, the ball tracking would start in the

next frame. This algorithm was successfully used in

videos of a basketball tournament, thus eliminating

the need of additional cameras as in multi-camera

estimation.

Another tracking technique, called mean-shift,

was used by (Kim, 2015) by detecting the colour and

edges of a curling stone.

On the other hand, (Yu et al., 2003) have used

trajectory-based optimization for the tracking of a

football in a broadcast video. In this case, the

algorithm does not evaluate if an object is a ball,

instead, it evaluates if a candidate trajectory is a ball

trajectory.

More advanced approaches include the use a

physics-based algorithm for predicting the ball’s

trajectory in 3D. This algorithm exploits the 2D ball

trajectory, along with the court lines, to reconstruct

the 3D trajectory and compute the shooting location.

This technique has been successfully used in videos

ICINCO 2018 - 15th International Conference on Informatics in Control, Automation and Robotics

512

of basketball (Chen et al., 2009) and volleyball (Chen

et al., 2011) matches.

In sum, the implementation of object tracking

techniques, such as the one presented above, in sports

can offer a panoply of advantages, as improving the

spectators’ experience, enhancing the training

process of professional athletes, providing valuable

information for the referee’s decisions, presenting

richer statistical analysis (Wang et al., 2006) or, as the

authors of this paper propose, motivating physical

activity.

2.1 The Boccia Game

Boccia is a strategy-based precision ball game that is

not as physical demanding compared to other sports.

The player only has to perform a throwing movement,

which allows the participation of individuals with

different levels of impairments.

Boccia can be played individually, by pairs or by

teams. In the latter, which is the case considered in

this paper, players are divided into two teams, the red

and the blue. Each team features three players and

each of them receives two balls of the respective team

colour. Briefly speaking, the main goal of the game is

for the players to place their team balls as close as

possible to a white ball, called the jack.

According to the BISFed International Boccia

Rules (BISFed, 2017), each Boccia match is divided

into six “ends”. Each end finishes whenever both

teams have thrown all of their respective balls and the

game score is noted as it follows:

 The team that placed the ball closer to the

jack scores one point for each ball placed at a

shorter distance from the jack than the closest

opponent’s ball to the jack.

 If two or more balls of different colour are

equidistant from the jack, each team scores a

point per ball.

The sum of the points of all the six ends determine

the winning team. If both teams finish with the same

number of points, further ends will be played until the

tie is undone.

In previous work developed by the authors of this

paper (Calado et al., 2018), the game score was

computed according to the aforementioned rules.

Although these are the official game rules, Boccia can

be easily readapted for a nursing home environment.

In the nursing home where the system will be tested,

the game rules have been changed by the caregivers

in order to simplify it as much as possible. This has

been done due to some of the residents having serious

cognitive disabilities.

Having this in account, during each end, each

time a team throws a ball and hits the jack, this team

scores one point. Similar to the official rules, the

winning team is determined by the total of points

from the six ends. In case of a tie, the team that has

the ball closest to the jack after the sixth end finishes

wins the game.

3 PROPOSED SYSTEM

In this section it is described the proposed system that

allows the graphical user interface (GUI) to display

the real-time score of a Boccia game. The score was

calculated using rules suited to the specific features

of the nursing home, as previously described.

3.1 Experimental Setup

The system’s architecture (Figure 1) consists of a

camera, a computer and a graphical user interface

(GUI).

Figure 1: System Overview.

The RGB webcam used was the Microsoft

LifeCam VX-1000, with a resolution of 0.3 MP (640

x 480), a field of view (FOV) of 50º and a maximum

frame rate of 30 fps.

3.2 System Description

Figure 2 shows the block diagram of the developed

system.

Boccia Court Gaussian Blurred

Colour

Segmentation

Display GUI

Frame Filtered Frame

Ball DetectionCalculate Score

Balls

Data Package

(colour,x,y)

Figure 2: Block diagram of the developed system.

The webcam captures the Boccia court and

provides a stream of frames, at 30 fps. Each

individual frame is analyzed to detect the ball

Boccia Court Analisys for Real-time Scoring

513

position. Initially, the frames go through a pre-

processing stage, where a Gaussian blurred filter is

applied to reduce image noise.

The computer runs an application developed in

Python programming language that implements a

colour segmentation technique to detect the balls

positions. To detect all the three ball colours (blue,

red and white), a colour mask was defined in HSV

(Hue Saturation and Value) space. In order to be

dependent on the light conditions, a sensitivity

margin was implemented, by trial and error, to cover

the desired range of colours.

Next to the colour segmentation, a few computer

vision techniques such as erode and dilate were

performed as well as the evaluation of some ball

metrics, such as radius and circularity in order to

create a set of valid balls. Then, each ball was

categorized with its colour, center, radius, x and y

positions.

When the balls were correctly detected, the

algorithm computed the distance between each of the

blue and red balls and the jack. If this distance was

less than a threshold value and the jack ball had

moved, it would be considered a valid point.

Finally, a data package was created with the

current score and sent to the developed GUI to be

displayed.

3.3 Communication

The communication between the algorithm that

calculates the real-time score and the GUI is done

using JSON format over the TCP/IP protocol.

An example of a JSON message to send/receive

the score can be seen below:

{

“instruction”: “SCORE_INFO”,

“red_points”: <[0, 3]>,

“blue_points”: <[0, 3]>

}

3.4 Graphical User Interface

The Qt framework was used for the development of

the GUI, which can be seen in Figure 3.

The GUI was developed with the goal of being

intuitive and to be easily understood. It can be

described in three parts: left side, centre and right

side, where the left side concerns the red team and the

right side the blue team.

On both the left and right sides, it is possible to

see – on top – the score of the current game scenario

– on the middle – the registry of the scores from the

six ends – and on the bottom – the total number of

points accumulated during a game.

Figure 3: Developed GUI in Portuguese language. Dashed

in blue – current score; dashed in green – number of the end

being played; dashed in yellow –registry of the score from

the six ends; dashed in white – total score.

On the centre, the number of the current partial

being played is displayed. There are also two buttons,

one to increment the partial and another to reset the

system.

4 PRELIMINARY RESULTS

The proposed system was tested in a nursing home,

where the elders were familiar with the Boccia game

with adapted rules, in S. Torcato, Guimarães,

Portugal. As referenced by the caregiver, these rules

state that, at beginning of the game, the jack is placed

at an arbitrary distance from the players, according to

the individual’s impairments. Moreover, a team

scores when a ball is thrown and collides with the jack

ball.

In previous work, the authors have recorded

several videos from different perspectives of

simulated game situations, along with the actual

playing of four ends. From the latter, only three of the

videos contained situations where a coloured ball

collided with the jack. Thereby, the algorithm was

tested in each of these three recordings and the

computed score was compared with the real score

corresponding to that situation.

Considering the three videos, the algorithm

computed the score from two of them correctly. In the

last video, the score was calculated incorrectly due to

an issue with the camera’s perspective, which caused

the algorithm to assume falsely that a collision had

occurred. In this case, a collision between a red and

blue ball caused the red ball to jump in the air and

partially occlude the jack, thus misleading the

algorithm to compute a false result by awarding one

point to the red team. In spite of the small number of

cases, the algorithm showed promising results.

Figure 4 depicts a game situation, contained in

one of the videos, where a blue ball collides with the

ICINCO 2018 - 15th International Conference on Informatics in Control, Automation and Robotics

514

jack. As it can be seen in Figure 4 (a), the blue ball

has a yellow arrow signalling the motion direction

towards the jack. After the collision, Figure 4 (b)

displays the correctly computed game score.

(a)

(b)

Figure 4: Example of one of the recorded game situations.

In situation (a), no previous scoring events have occurred,

and the blue ball is going to hit the jack. Later, in situation

(b), the blue ball had collided with the jack ball and the

score of one point for the blue team is computed correctly.

5 FINAL REMARKS

In this paper, the authors proposed a system based on

a computer vision algorithm that allows the automatic

computation of the game score during a Boccia match

and its display in a GUI. The purpose of this system

is to enhance the playing experience of the Boccia

game and to encourage the elderly to engage in

physical activity on a regular basis.

The computer vision algorithm was tested in three

videos, recorded in the nursing home, containing

Boccia gameplay with ball collisions. The obtained

results showed that all collisions were detected

correctly, except for one. False positives, such as this

one, can occur due to the limited 2D perspective,

which may induce the algorithm into assuming that a

collision occurred when the jack becomes occluded

by another ball or vice-versa.

Regarding ball detection, luminosity conditions

are also an issue because they can affect the colours

reproduced in the camera’s image, leading to false

positives, thus affecting the game score. The

detection based on colours, by itself, is also an issue

due to the need of manually adjusting parameters if

the recording environment is changed. A solution for

this issue could be the implementation of a calibration

method, hence making the system more flexible.

Regarding future work, the developed system is

going to be evaluated in a real-time scenario during a

Boccia match played by the resident elders of the

nursing home in S.Torcato in order to assess its

performance. This evaluation will be based on several

metrics, in a first phase, the algorithm will be tested

to reduce false positives and improve the detection

rate, and, in a second phase, a usability test will be

performed. The objective of performing such tests is

to infer if the elders can easily understand the

graphical display containing the score. It is also

important to comprehend the level of trust of the users

in the proposed system, along with their degree of

satisfaction. Regarding the promotion of physical

activity, as can been seen in serious games, in most

cases a rewarding system is implemented. This could

be done by simply including the photo/name of the

player on the GUI when he/she wins. It would also be

interesting to include in the GUI the photo/name of

the user while he/she plays.

The usage of machine learning for ball detection

and tracking is also a possibility to be explored in the

future. Such techniques could help improving the

system’s performance by having a more robust

detection rate and a more stable tracking.

ACKNOWLEDGEMENTS

This article is a result of the project Deus ex Machina:

NORTE – 01 – 0145 – FEDER - 000026, supported

by Norte Portugal Regional Operational Programme

(NORTE 2020), under the PORTUGAL 2020

Partnership Agreement, through the European

Regional Development Fund (ERDF).

REFERENCES

Arroxellas, R.D. de, Romano, R.G., Cymrot, R., Blascovi‐

Assis, S.M., 2017. Bocha adaptada: análise cinemática

do arremesso e sua relação com a realidade virtual. Rev.

Bras. Ciencias do Esporte 39, 160–167.

BISFed, 2017. BISFed International Boccia Rules (v.2).

Boccia Court Analisys for Real-time Scoring

515

Calado, A., Leite, P., Soares, F., Novais, P., Arezes, P.,

2018. Boccia Court Analysis for Promoting Elderly

Physical Activity. Submitted.

Chen, H.T., Tien, M.C., Chen, Y.W., Tsai, W.J., Lee, S.Y.,

2009. Physics-based ball tracking and 3D trajectory

reconstruction with applications to shooting location

estimation in basketball video. J. Vis. Commun. Image

Represent. 20, 204–216.

Chen, H.T., Tsai, W.J., Lee, S.Y., Yu, J.Y., 2011. Ball

tracking and 3D trajectory approximation with

applications to tactics analysis from single-camera

volleyball sequences. Multimed. Tools Appl. 1–27.

Direção-Geral da Saúde, 2017. Programa Nacional Para a

Promoção Da Atividade Física 2017. Direção-Geral da

Saúde.

Estabrooks, P. a, Carron, A. V, 1999. Group cohesion in

older adult exercisers: prediction and intervention

effects. J. Behav. Med. 22, 575–588.

Figueira, C., Silva, J., Santos, A., Sousa, F., Silva, V.,

Ramos, J., Soares, F., Novais, P., Arezes, P., 2017.

iBoccia: Monitoring elderly while playing Boccia

gameplay. ICINCO 2017 - Proc. 14th Int. Conf.

Informatics Control. Autom. Robot. 1, 670–675.

Gillespie, L.D., Robertson, M.C., Gillespie, W.J.,

Sherrington, C., Gates, S., Clemson, L.M., Lamb, S.E.,

2012. Interventions for preventing falls in older people

living in the community. Cochrane Database Syst. Rev.

Hawk-Eye [WWW Document], 2018. .

www.hawkeyeinnovations.com.

Huang, P.C., Pan, P.J., Ou, Y.C., Yu, Y.C., Tsai, Y.S.,

2014. Motion analysis of throwing Boccia balls in

children with cerebral palsy. Res. Dev. Disabil. 35,

393–399.

Instituto Nacional de Estatística, 2016. Inquerito Nacional

de Saúde 2014, Ine.

Kim, J., 2015. Curling Stone Tracking by an Algorithm

Using Appearance and Colour Features. Proc. World

Congr. Electr. Eng. Comput. Syst. Sci. 1–6.

Landi, F., Cesari, M., Onder, G., Lattanzio, F., Gravina,

E.M., Bernabei, R., 2004. Physical activity and

mortality in frail, community-living elderly patients. J.

Gerontol. A. Biol. Sci. Med. Sci. 59, 833–837.

Lee, I.-M., Shiroma, E.J., Lobelo, F., Puska, P., Blair, S.N.,

Katzmarzyk, P.T., 2012. Impact of Physical Inactivity

on the World’s Major Non-Communicable Diseases.

Lancet 380, 219–229.

Matthews, C.E., Chen, K.Y., Freedson, P.S., Buchowski,

M.S., Beech, B.M., Pate, R.R., Troiano, R.P., 2008.

Amount of time spent in sedentary behaviors in the

United States, 2003-2004. Am. J. Epidemiol. 167, 875–

881.

Pingali, G., Opalach, A., Jean, Y., 2000. Ball tracking and

virtual replays for innovative tennis broadcasts.

Proceedings-International Conf. Pattern Recognit. 15,

152–156.

Salguero, A., Martínez-García, R., Molinero, O., Márquez,

S., 2011. Physical activity, quality of life and symptoms

of depression in community-dwelling and

institutionalized older adults. Arch. Gerontol. Geriatr.

53, 152–157.

Schutzer, K.A., Graves, B.S., 2004. Barriers and

motivations to exercise in older adults. Prev. Med.

(Baltim). 39, 1056–1061.

Silva, Vinícius; Ramos, João; Soares, Filomena; Novais,

Paulo; Arezes, Pedro; Sousa, Filipe, Silva, Joana;

Santos, A., 2018. iBoccia: A Framework to Monitor the

Boccia Gameplay in Elderly. Lect. Notes Comput. Vis.

Biomech. 27.

Stessman, Jochanan; Hammerman-Rozenberg, Robert;

Cohen, Aaron; Ein-Mor, Eliana; Jacobs, J.M., 2009.

Physical activity, function, and longevity among the

very old..pdf 169, 1476–1483.

Tsai, Y.S., Yu, Y.C., Huang, P.C., Cheng, H.Y.K., 2014.

Seat surface inclination may affect postural stability

during Boccia ball throwing in children with cerebral

palsy. Res. Dev. Disabil. 35, 3568–3573.

Wang, J. R.; Parameswaran, N., 2006. Survey of Sports

Video Analysis: Research Issues and Applications. Sch.

Comput. Sci. Eng. Univ. New South Wales. 113, 115–

118.

Wang, I.-L., Huang, Y.-C., Lee, R.-H., Sun, Y.-M., 2017.

Development and analysis of the boccia scoring system.

Int. J. Comput. Inf. Syst. Ind. Manag. Appl. 9, 80–86.

Westhoff, T.H., Franke, N., Schmidt, S., Vallbracht-Israng,

K., Meissner, R., Yildirim, H., Schlattmann, P., Zidek,

W., Dimeo, F., Van Der Giet, M., 2007. Too old to

benefit from sports? The cardiovascular effects of

exercise training in elderly subjects treated for isolated

systolic hypertension. Kidney Blood Press. Res. 30,

240–247.

WHO, 2015. Factsheets on Health-Enhancing Physical

Activity in the 28 European Union Member States of

The WHO European Region 1–27.

WHO, 2010. Global recommendations on physical activity

for health. Geneva World Heal. Organ. 60.

Wu, L., Meng, X., Liu, X., Chen, S., 2006. A new method

of object segmentation in the basketball videos. Proc. -

Int. Conf. Pattern Recognit. 1, 319–322.

Yu, X., Xu, C., Leong, H.W., Tian, Q., Tang, Q., Wan,

K.W., 2003. Trajectory-based ball detection and

tracking with applications to semantic analysis of

broadcast soccer video. Proc. Elev. ACM Int. Conf.

Multimed. - Multimed. ’03 pp, 11.

ICINCO 2018 - 15th International Conference on Informatics in Control, Automation and Robotics

516