Mii School: New 3D Technologies Applied in

Mii School: New 3D Technologies Applied in Education to Detect Drug Abuses and Bullying in Adolescents José Alberto Carmona1, Moisés Espínola2, Adolfo J. Cangas1, and Luis Iribarne2 1

Department of Clinical Psychology, University of Almería, Spain 2 Applied Computing Group, University of Almería, Spain {jacarmona,moises.espinola,ajcangas,luis.iribarne}@ual.es

Abstract. Mii School is a 3D school simulator developed with Blender and used by psychology researchers for the detection of drugs abuses, bullying and mental disorders in adolescents. The school simulator created is an interactive video game where the players, in this case the students, have to choose, along 17 scenes simulated, the options that better define their personalities. In this paper we present a technical characteristics description and the first results obtained in a real school.

1 Introduction: Virtual Reality in Education The enormous development that new technologies, such as virtual reality, have experienced in the last years has allowed the emergence of a growing number of studies focusing on applying these technologies in the educational environment. In this sense the use of learning models through virtual reality programs has several advantages over traditional methods of teaching, such as the possibility of multimodal learning or the facilitation of the skills transference processes learned from one context to another (Freitas & Neumann, 2009). Thus, as some authors indicate, the learning process using virtual reality takes place in a more practical and less limited form than in traditional texts based approaches. Hence particular importance is the use of video games for performance improvement in the learning process, since they add several advantages to this process, as sometimes has been studied in psychology research. Specifically, a study by Kim, Park and Baek (in press) proves that video games, used in conjunction with certain meta-cognitive strategies, produce an improvement not only in academic learning, but also game itself performance. In turn, virtual environments that recreate school contexts have been developed for different purposes, such as virtual classes recreation in order to assess and rehabilitate attention deficits in young people (Rizzo et al. 2001) or the called virtual schools which are presented as useful learning tools in the education field, because they give enormous benefits using virtual environments for teaching, such as improving skills and academic results or expanding education access. However there is little research related to the study of these virtual schools (Barbour & Reeves, 2009). On the other hand, there are few studies focused on the application of virtual reality to the field of students mental disorders assessment or treatment in the educational environment. There is a software called FearNot! oriented to treat bullying behaviours inside a school context. This program recreates, through 3D simulation, violent situations in school environment with the objective of consequences learning of these M.D. Lytras et al. (Eds.): TECH-EDUCATION 2010, CCIS 73, pp. 65–72, 2010. © Springer-Verlag Berlin Heidelberg 2010

66

J.A. Carmona et al.

behaviours and thus the student deletes these types of behaviour in school (Zoll, Enz, Schaub, Aylett & Paiva, 2006). In addition, another researches have study the treatment of anxiety in exams and school phobia through the recreation of the significant situations in which these problems appear by means of virtual reality (GutiérrezMaldonado, Alsina-Jurnet, Carvallo-Becíu, Letosa-Porta y Magallón-Neri, 2007). In this paper we present a new virtual reality tool called Mii School, a 3D school simulator developed with Blender (Hess 2007) and used by psychology researchers for the detection of drugs abuses, bullying and mental disorders in adolescents. The school simulator created is an interactive video game where the players, in this case the students, have to choose the options that better define their personalities along 17 scenes simulated. Different quality versions of Mii School have really been created, so the execution of the program can be adapted to the technical characteristics of each specific computer as closely as possible. So if Mii School is executed on a latestgeneration PC, it can use the highest-quality version to enjoy better viewing (XGA 1024x768 resolution). However, if the graphic card can not reproduce this visual quality with sufficient fluency, other lower-quality versions can be executed (SVGA 800x600 or VGA 640x480 resolutions). In section 2 we describe the different scenes represented in Mii School, and later in section 3 we explain the Mii School development process. In section 4 we present the satisfaction results obtained after applying this tool to several students, and we finally show the future works in section 5.

2 Scenes Description of Mii School During the execution of Mii School the student watches a total of 17 interwoven scenes studying different aspects of his behavior related to bullying, drug addiction, family life, capacity for attention in class and integration in social groups.

Fig. 1. Screenshots of bullying and drugs scenes in Mii School

Mii School: New 3D Technologies Applied in Education to Detect Drug Abuses

67

There are a total of 5 scenes that study bullying. In some of them, the student is bullied by his schoolmates and in others he becomes the bully, in this way bullying can be studied from several perspectives. Some concrete cases are also studied, for example, the reaction of the student to the explicit violence of a physical aggression in the schoolyard to see if he is mediator or violent. For each bullying scene, the student can select one of a series of choices that always follow a general pattern: feeling indifferent to the bullies, protesting to them, responding ironically, running away in fear, facing up to them or feeling ashamed. The student’s relationship with his parents is also studied in three scenes: one scene checks the father’s behavior, another one the mother’s attitude and a third scene both behavior . In the first two scenes, the student gets home much later than the hour agreed upon and checks whether the father or the mother scold him, threaten him or, on the contrary, are indifferent to his undisciplined behavior. In the third scene, the student gets home after school and feels anxious because he has problems with his studies and the simulation checks whether his parents become involved in his problems or are indifferent to them. During the 6 drug addiction scenes, the drug offered becomes gradually more dangerous. In the first scene, the student is tempted by his schoolmates to smoke in the school playground. In the second scene, he is invited to drink alcohol while eating a pizza at a friend’s house. Afterwards, he is offered a joint of marihuana in a park. In the last two scenes, the risk is upped further when his friends offer him much more dangerous drugs, such as cocaine or ecstasy (MDMA). The choices that the student can choose in the drug addiction scenes also follow a general scheme: usually use, refuse to try the drug, advice friends to stop taking it, use occasionally or leave because he feels uncomfortable. There are other scenes where personality-related problems and the student’s mood, attention in class, beliefs and integration in social groups are checked. The following table shows the most important metrics in the 17 Mii School simulated scenes. Table 1. Some interesting metrics of Mii School simulated scenes SCENE

THEME

STAGE

DURATION

NºPEOPLE

NºCAMERAS

NºOPTIONS

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17

Bullying Drugs (tobacco) Bullying Drugs (alcohol) Attention in class Social integration Bullying Drugs (marihuana) Drugs (cocaine) Relation with father Relation with mother Relation with parents Drugs (general) Bullying Drugs (ecstasy) Bullying Beliefs

Courtyard Courtyard Classroom Home Classroom Courtyard Courtyard Park Home Home Home Home Park Courtyard Park Courtyard Various

1:36 s 1:12 s 1:09 s 1:09 s 51 s 42 s 54 s 1:12 s 1:06 s 1:03 s 1:03 s 1:03 s 42 s 1:15 s 1:06 s 1:21 s 39 s

4 3 6 2 5 5 5 4 2 2 2 3 3 2 2 4 12

6 5 4 4 3 4 3 4 3 5 5 6 4 5 4 6 8

6 5 6 5 5 4 6 5 5 4 4 4 4 9 5 6 8

68

J.A. Carmona et al.

As may be observed in the table, the number of choices that can be made by the user usually coincides in the scenes that share the same theme.

3 Mii School Development Methodology We have chosen Blender as Mii School development tool because this software incorporates all the functionality needed: 3d design, advance illumination and texturing, characters animation, blender game engine, programming with Python, etc. Besides, Blender has got the most important property of a most wanted software: it is open source. Mii School development process with Blender was divided into five different stages: 3D design, texturing/illumination, animation, Python programming and Blender Game Engine configuration. 3.1 3D Design During this first stage, the 3D meshes were created with Blender to represent the virtual scenarios and characters that appear in the scenes (Roosendaal and Selleri, 2005). The Mii School simulation takes place into different scenarios: the school playground, the classroom, the main character’s home, a park... Nevertheless, each scenario is divided in several zones (for example, the home has several rooms, the playground has an area with benches, another for sports, etc.) so each zone is used for a different scene and the user does not get the impression that the scenarios are repeated. A total of 30 characters were designed in Blender: several male and female students of different races and social conditions, the teacher who is teaching in the classroom, the student’s parents, etc. All of the Mii School 3D models were created using basic and advanced design techniques included in Blender: extrude, split, merge, etc. and finally, smoothing filters were applied for smoother surfaces without overly increasing the number of polygons (set smooth option). 3.2 Texturing/Illumination Texturing was also done during this first stage using UV mapping techniques and Blender materials editors, applying good quality images to the 3D models to increase realism of the scenes and characters. For proper illumination of the scenes, a sun type global lighting illuminates above the main student, on whom the simulation action always focuses, so that the objects and characters around him are always properly illuminated. Some spotlights were also used to increase illumination of certain places. 3.3 Animation Once the 3D models were designed, we proceeded to the animation stage by first creating skeletons associated with the 3D characters meshes, and then capturing bones movement using a series of intermediate poses with the aid of a technique called inverse kinematics (Hess 2008). Characters animations were divided into two large groups: body and facial ones. Body animations affects the character’s whole body. We implemented several types


69

of body animation in which the characters perform some action: walking, sitting, hitting, threatening, fighting, smoking, drinking, etc. In order to increase the expressiveness and realism of facial animations, the number of polygons and level of detail have been increased noticeably in eyes and lips, areas that most influence the facial gestures. So if one character is threatening another in a certain scene, his face expresses aggressiveness, or if on the contrary a character feels threatened, his face shows fear. 3.4 Python Programming When the characters were created and the corresponding animations implemented, we went to the coding stage using the Python programming language (Sherrod 2008). Over 6.000 source code lines have been implemented in Python to simulate the action of the 17 scenes. This source code is the Mii School kernel, it specifies all the details necessary to compose the scenes, and this kernel also captures all of the information provided by the student. The source code specifies aspects such as which characters intervene in each of the scenes or which scene is going to be developed at this moment. In order to achieve a better performance during program execution, along the simulation of a scene only the characters and objects that are going to be viewed by the camera are shown, and the rest are hidden. The source code that we have programmed in Python follows the instructions of a deterministic finite automata to find out in what state of simulation Mii School is at any given time and to set up the escene: activate the corresponding camera, the body animation and facial expressions of the character who is speaking at the moment, etc. The following text shows part of the source code for the first scene in Mii School. In this code it is specified the person speaking in this moment, his basic body animation, the facial expression (mouth and eyebrows), the active camera and the dialogue that is visualized on the screen. Finally, the complete scene is activated by sending all of the data to the Blender Game Engine. !"# $ % & #'"# $ ( # ! # ( # )&' $ ! $$* '! + $$,,, ( + (. ** *% &"* *)/ 0

Fig. 2. Part of Python source code in Mii School

3.5 Blender Game Engine The Mii School kernel is interpreted by the Blender Game Engine. In order to simulate Mii School, the Blender Game Engine not only executes the Python source code that we have developed, but also takes into account a series of events associated with

70

J.A. Carmona et al.

the dynamic objects in the scenes, especially characters (both their 3D mesh and the skeleton) and cameras (with their corresponding dialogue box, which also has associated events). Each event is divided into 3 parts: the sensor (that receives information, for example, pressing a key or a specific signal), the actuator (which performs actions on objects, like animating a character or moving a camera) and the controller (which connects the sensor to the actuator). In the Mii School project, there are over 900 basic objects with associated events (some of them are complex objects such as a character’s skeleton or its 3D mesh, and another are simpler, such as the dialogue box for a specific camera), and each object has an average of 10 events of its own that regulate their functioning (for example, in the cameras an event can indicate a movement and in a character the activation of an animation). The Blender Game Engine also takes another secondary details into account, such as in which scenes the main character’s cigarette or bottle should be seen or hidden, or what color should be the conversation box depending on the character who is speaking at that moment. To summarize, during the execution of Mii School the Blender Game Engine must synchronize around 10.000 different events in real time, activating only those that are necessary right at the moment of the simulation and discarding the rest. Apart from this work, it also is in charge of rendering and illuminating the scenes in real time, as well as capturing information introduced by the user.

4 Satisfaction Results The total simulation of Mii School 17 scenes lasts approximately 25 minutes. The student can choose his sex at the beginning of the simulation to personalize the animations to his own gender. All the information introduced by the student, as well as the choices selected in each one of the scenes and other data of interest, is stored in a web page format for the later viewing and analysis by the psychology researchers. Background music has been added in the scenes and audio in the conversations to achieve more realism. Proper movement of the cameras during the transition of scenes is also carefully made so the simulation has the quality of real movie. This section shows the results from a first application of Mii School in a real educational environment. The showed data specifically mentions the satisfaction degree of participants with regard to matters concerning characteristic that our software has got. In the pilot study, 65 students have participated: 30 men and 35 women. These participants belong to the first, second, third and fourth year of the Spanish Secondary School (E.S.O.) of a concerted public education school in Almería (Spain). The age of the participants was between 13 and 16 years old, with a mean age of 14.63 (SD: 1.14). The following data was taken with the help of a satisfaction questionnaire developed specifically for this study that evaluates general issues concerning the general Mii School features using a scale of Likert, with 5 response options, from “strongly disagree, never, nothing” to “strongly agree, quite, always”, including the neutral opinion “do not know, no opinion”. After evaluating the student´s responses, we can see high percentages of participants satisfied with Mii School characteristics. Thus all the issues raised, except one, have a positive value of 80-90% of participants (in regards like identifying the subject of the scenes, if appropriate instructions


71

have been introduced, or time of recreated situations). Only the item related with the Mii School program familiarity had a slightly lower value (specifically the 66.2%). Table 2. Satisfaction of Mii School users ITEMS

% students

It was easy to identify the subject (drugs, bullying) from each scene

80%

The instructions received by the program are enough to understand how it works The scenes information (dialogues, interactions, etc) was enough to understand the situations described The playing time of each scene was enough to understand its content.

89,2% 87,7% 90,8%

The responses playing time has been enough to recognize the options available

89,2%

I found family the use of Mii School,given that I am used to these technologies

66,2%

Mii School situations are appropriate for evaluating the issues of each scene.

89,2%

In the same study we compared the effectiveness of drugs detection and bullying behaviours with Mii School using several traditional pencil and paper assessment questionnaires. In general, Mii School was more sensitive to the detection of bullying behaviours and drugs consumption than paper test (Carmona, Espínola, Cangas & Iribarne, in press).

5 Future Works As future works we want to add to Mii School several modules that will allow this software not only the drugs consumption detection, but also the drugs abuse prevention. Besides Mii School will be translated into several languages, and we are thinking about an online version of Mii School.

Acknowledgments This work was financed with a research project from the SPANISH MINISTRY OF HEALTH AND CONSUMPTION (NATIONAL PLAN ON DRUGS, ref. 2007/063) awarded to the third author.

References Barbour, M.K., Reeves, T.C.: The reality of virtual schools: A review of the literature. Computers & Education 52, 402–416 (2009) Carmona, J.A., Espínola, M., Cangas, A.J., Iribarne, L.: Mii-School: A 3D video game for the early detection of abuse of substances, bullying and mental disorders in adolescents. European Journal of Education and Psychology (in press) Freitas, S., Neumann, T.: The use of ‘exploratory learning’ for supporting immersive learning in virtual environments. Computers & Education 52, 343–352 (2009)

72

J.A. Carmona et al.

Gutiérrez-Maldonado, J., Alsina-Jurnet, I., Carvallo-Becíu, C., Letosa-Porta, A., MagallónNeri, E.: Aplicaciones clínicas de la realidad virtual en el ámbito escolar. Cuadernos de medicina psicosomática y psiquiatría de enlace 82, 32–51 (2007) Hess, R.: The essential Blender: guide to 3D creation with the open source suite Blender, 376 pages (2007) Hess, R.: Animating with Blender: how to create short animations from start to finish, 368 pages. Focal Press (2008) Kim, B., Park, H., Baek, Y.: (en prensa). Not just fun, but serious strategies: Using metacognitive strategies in game-based learning. Computers & Education Rizzo, A.A., Buckwalter, J.G., Bowerly, T., Humphrey, L.A., Neumann, U., Rooyen, A., et al.: The virtual classroom: a virtual reality environment for the assessment and rehabilitation of attention deficits. Revista Española de Neuropsicología 3, 11–37 (2001) Roosendaal, T., Selleri, S.: The official Blender 2.3 guide: free 3D creation suite for modeling, animation and rendering, 784 pages. No Starch Press Publish (2005) Sherrod, A.: Game graphics programming, 645 pages. Charles River Media Publish (2008) Zoll, C., Enz, S., Schaub, H., Aylett, R., Paiva, A.: Fighting bullying with the help of autonomous agents in a virtual school environment. Dissertation, Trieste, Italy (2006)