سیستم کامپیوتری هنر درمانی برای طراحی خانواده سینتیکی

This study delineates the development of the Computer Art Therapy System for the Kinetic Family Drawing (CATS_KFD) which uses given patterns instead of free drawings. The CATS_KFD consists of four stages. The system provides clients with questionnaires on their family relationships and color preferences, and their answers constitute a fact base. The system provides the client with various patterns of family members and backgrounds which possibly compose a KFD, and the client selects a few among them; manages, expands, or contracts; and colors them. The system evaluates elements and detects changes of evaluations in the KFDs, and this evaluation and detection of changes also constitute the fact base. The system interprets one or several KFDs by invoking the knowledge in a knowledge base corresponding to the facts in the fact base. The knowledge base, which consists of knowledge related to the KFD is continuously accumulated and updated. The system automatically provides art therapists not only with an objective evaluation of the elements, but also with information about their client's psychological status and treatment process. It is expected that the CATS_KFD with patterns would inherit and complement the validity of the traditional KFD with free drawings.

This study delineates the development of the Computer Art Therapy System (CATS) which operates in real-time via the internet for the Kinetic Family Drawing (KFD) developed by Burns and Kaufman (1972). We designate the system as the CATS_KFD. First of all, it should be admitted that this study does not use KFDs that are freely drawn, but only KFDs that are composed of given patterns. Despite the remarkable progress of computer technology, at present, the computer is unable to identify forms in free drawing and, thus, the actions of forms and the spatial relationships among them. Thus, for the system to identify the subjects in a drawing, the system prepares in advance various kinds of patterns of family members and backgrounds which will possibly compose a KFD, and provides them to the client, and the client selects a few among them; arranges, expands or contracts; and colors them. Then, by inbuilt functions of the computer and some existing CATSs for art assessments, the system can identify the forms and evaluate the placement, size, distances between forms, and various other elements. We examined all of the elements which are possibly considered for evaluation in the KFD as much as we can in various literature, including Burns and Kaufman (1972), McPhee and Wegner (1976), Myers (1978), Reynolds (1978), elements in the Descriptive Assessment of Psychiatric Artwork (Hacking, 1999), the Diagnostic Drawing Series (DDS) (Cohen, 1986/1994), the Person Picking an Apple from a Tree (PPAT) (Gantt, 1990), the Computer Color Related Elements Art Therapy Evaluation System (C_CREATES) (Kim, Bae, & Lee, 2007), and structured mandala (Kim, Kang, & Kim, 2009). After consolidating those elements which were the same but with different names, we found that there were 127 elements. For the evaluation of these elements, there are many positive reports/papers showing high inter-rater reliability. Mostkoff and Lazarus (1983) mentioned the possibility of creating an objective evaluation system, and McPhee and Wegner (1976), Cummings (1980), and Elin and Nucho (1979) reported high inter-rater reliability. However, some elements have to be evaluated by the human rater's intuition, judgment, and subjective determination and, thus, some inconsistencies are inevitable. Even for some elements using certain selected objective criteria, such as rulers or grids made of tracing paper marked off in millimeters, this process is time consuming and there still remains some inaccuracy. Several CATSs have been developed for art assessment which provide absolute reliability in the evaluation of some elements in drawings, viz. the CATSs for the evaluation of color-related elements, C_CREATES (Kim, Bae, et al., 2007), determination of placement (Kim, Kang, & Kim, 2008), judgment of main color (Kim, 2008), evaluation of variety of color (Kim & Hameed, 2009), and evaluation of elements in structured mandala (Kim, Kang, et al., 2009). By employing these computer systems and inbuilt functions of the computer, the CATS_KFD can automatically, accurately, precisely, quantitatively, and objectively evaluate the 127 elements in the CATS_KFD, thus achieving absolute consistency in its assessment. As a subject of the CATS, we chose the KFD developed by Burns and Kaufman (1972), since it is one of the most widely used art therapy tools and its reliability and validity have been proven. We expect the CATS_KFD with given patterns to inherit the applicability, usefulness, and validity of the traditional KFD with free drawings. Burns and Kaufman argued that the KFD reflects emotional disorders faster than other methods do and offers art therapists information not only about the adaptive and defensive functions of children, but also about their family dynamics. There are plenty of literatures reporting the validity of the KFD in applying it for various purposes. For example, the KFD recognized cases of child abuse or neglect, or distinguished children who are suffering from child abuse or neglect from normal children in the studies of Schornstein and Derr (1978), Hackbarth, Murphy, and McQuary (1991), Reddy, Bhadramani, and Samiullah (2002), and Veltman and Browne (2003). We omit its validity in other applications. On the other hand, it is well known that while many studies found positive results using the KFD, other studies were not as confident of its validity. Cummings (1980) suggested that KFDs may be sensitive to the transition in children's personality states and Mostkoff and Lazarus (1983) asserted that the KFD is sensitive to children's mood changes, and therefore it may not be an accurate measure of personality traits or characteristics. There are even some studies reporting contradictory results. Monahan (1986) noted that more psychological symptoms are found in the KFDs of high-achieving children than those of lower-achieving children. Their results failed to support the hypotheses of Burns and Kaufman (1972), who used the KFD to determine the extent of a child's adjustment. We omit other studies reporting questionable validity or contradictory results. For this problem, we expect that the CATS_KFD could be a helpful tool as an experiment for the validity of knowledge, which will be discussed in the later section of discussion and conclusion. The CATS_KFD consists of four stages. In the first stage of the questionnaires, in order to obtain information about the client's environment and history, which cannot be obtained from drawings, the system in this study provides two questionnaires regarding the client's family relationships and color preferences, whose answers constitute a fact base (FB). In the second stage of KFD composition and coloring, the client selects a few among various given patterns, arranges, rotates, expands or contracts, and colors them. In the third stage of analysis and evaluation of the elements in the composed KFD, the system evaluates the elements in a KFD or detects changes of evaluation in several KFDs by using inbuilt functions of the computer and some existing CATSs for art assessments. The results of the evaluation and detection of changes also constitute an FB. In the final stage of interpretation, the system provides information about the client's psychological status or treatment process by invoking the knowledge in the knowledge base (KB) corresponding to the facts in the FB. The KB, which consists of the expertise and experience of experts and literatures, is continuously accumulated and updated. The system developed so far is currently only a prototype model. However, all of the features have been designed. The procedure of the CATS_KFD that is operated in real-time on the internet is as follows: The user receives an explanation of how to operate the system, which takes 5–10 min. The system offers questionnaires and receives the user's answers (5–10 min). With appropriate questionnaires, various information about the client's environment and history can be obtained. Depending on the answers, the system offers various types of patterns for each family member and background. The user selects a few among these patterns, arranges them, and expands or contracts their sizes (15–20 min). In this procedure, the system can evaluate the client's sense of balance. Then, the user colors the patterns on the computer using computer brushes with three thicknesses (20–25 min) or prints the completed KFD with the patterns on an A4-sized paper and then colors them using crayons or markers, makes a file copy by means of a scanner or digital camera, and sends it to the system. In this client's coloring work, the system can evaluate three elements, viz. the degree of concentration, completeness, and accuracy. The system evaluates all of the elements in a KFD or detects changes of evaluations in several KFDs, and then interprets and informs the clients of the results (less than a minute). The system can provide art therapists with various valuable information about element evaluation such as the number of colors used, the area painted, family members missing, sizes of family members, distances between family members, etc., by evaluating the various elements in the KFDs with patterns. Also, it can aid art therapists in their decision making on a client's psychological status or treatment process by interpreting one or more KFDs. We hope this kind of interdisciplinary work incorporating art therapy and computer science can not only promote the use of various art therapy tools, but also pave the way for the development of new technology for the analysis and application of art therapy, as well as its theoretical understanding. In the following four sections, the four stages of the CATS_KFD procedure are explained with case examples. In the final section of discussion and conclusion, the contribution of this interdisciplinary study to the field of art therapy and its possibilities are mentioned.

We computerized the KFD by using patterns instead of free drawings. The prototype computer system, CATS_KFD, can automatically provide the art therapists with accurate, precise, objective, quantitative, and consistent evaluation with 100% reliability for the 124 elements in hundreds of KFDs in a minute by using inbuilt functions of the computer and several CATSs for art assessments. Also, the CATS_KFD aids the art therapists to make their decision in art therapy treatment of clients by interpreting one or several KFDs. It is expected that the CATS_KFD with patterns would inherit the original applicability, usefulness, and validity of the traditional KFD with free drawings. The system can evaluate the sizes of forms, not only in terms of their length, but also in terms of their area. Moreover, the CATS_KFD evaluates the client's sense of balance in the process of composing a drawing, and the completeness, accuracy, and degree of concentration in the coloring patterns, which are elements newly added to the traditional KFD. The CATS_KFD can be used as an experiment tool to set the criteria of the traditional KFD for evaluating the interpretation. For example, it could find and specify how small the resulting picture should be in the assessment to judge as low self-esteem, or how large the resulting distance between family members should be to see as a sign of a child's psychological sense of alienation. Since the CATS_KFD offers a new, easy, and interesting way of composing a drawing and coloring, compared to the traditional KFD, it may be more appropriate for clients who have severe developmental, psychological, or physical disorders. Also, the CATS_KFD could offer emotional stabilization and the improvement of the client's psychological state during the work of coloring given patterns, which were reported in coloring structured mandalas by Curry and Kasser (2005). There is a huge amount of expertise, experience, case studies, and results of experts who are world renowned authorities in art therapy, reports, papers, books, and literatures. Millions of items can be stored in a KB as a kind of dictionary, and from this, the CATS_KFD can retrieve each item in a second. In conclusion, as a worldwide authority of art therapy, the CATS_KFD can not only interpret a single KFD, but also detect any changes, trends, or process of art therapy treatment from a series of KFDs. In doing this, the system never shirks its duty, is never fatigued with its responsibility, and never makes a mistake of omission. We hope the CATS_KFD can be developed as a standardized instrument. With its reliability, validity, and expert system functions, the system can be a new methodology to be extended to the Human Figure Drawing (HFD) (Koppitz, 1968), Draw A Person (DAP) (Neglieri, McNeish, & Bardos, 1991), House-Tree-Person (HTP) (Buck, 1987), PPAT (Gantt, 1990), etc., with patterns until computer technology progresses to the point where it is able to identify the forms and symbols in a free drawing. It has sometimes been reported that the usefulness of projective techniques such as the KFD, HFD, DAP and HTP and the interpretation of art is in doubt, due to their lack of reliability and highly questionable validity. This might be due to the fact that some elements are too complex for raters to evaluate perfectly and that people and art are too complex for drawings to embody perfectly a given diagnosis (Cohen, 1986/1994). In particular, the interpretation of children's drawings is confounded by maturational issues (Kahill, 1984). However, we cannot deny the contribution and usefulness of projective techniques. From the viewpoint of artificial intelligence, the field of art therapy is a so-called Ill-Structured Paradigm (ISP) where numerous studies and methods have produced results that are diverse, lacking in consistency and sometimes even contradictory (Kim, 2008). There are many other important disciplines that are associated with the ISP, including clinical diagnosis, weather forecasting, and demand forecasting. Usually, there are several typical problems in the ISP: first, which item should be included in the KB and which item should not; second, some items have very high validity while others have much less validity; third, the consideration of individual situations. The approach of expert system in artificial intelligence is known to be a very useful solution to these kinds of problems. For the first problem, the system may be used in experiments for the validity of knowledge. For the second problem, the system can give some sort of validity measure to each item. For example, 0 for no evidence, 1 for weak evidence, 2 for possibility, 3 for moderate evidence, 4 for strong evidence, and 5 for absolute evidence. Also, the Bayesian network (Charniak, 1991) approach could be adopted. For the third problem, the system can consider a client's individual characteristics, environment, background, history, etc., with suitably designed questionnaires and the exceptional memory capability of computer. There have been studies on the application of expert systems to art therapy: the expert system approach (Kim, Ryu, Hwang, & Kim, 2006); reasoning process (Kim, Kim, Lee, Lee, & Yoo, 2006); framework of the KB (Kim, Yoo, Kim, & Lee, 2007). Also, several CATSs have been developed for art assessment and interpretation: the evaluation of color related elements in drawings, C_CREATES (Kim, Yoo, et al., 2007); determining unusual placement (Kim, Kang, et al., 2008); judging main color in a drawing (Kim, 2008); evaluation of variety of color (Kim & Hameed, 2009); evaluation of elements in mandala (Kim, Kang, et al., 2009); estimation of levels of psychological disorders (Kim, Betts, Kim, & Kang, 2009); mandala interpretation (Kim, Kim, & Kim, 2008). Our aim is to create a computer system which will promote the use of various art therapy tools and, furthermore, even pave the way for the development of new technology for the analysis and application of art therapy, as well as its theoretical understanding. There remains a tremendous amount of further study to be performed. We suggest that these interdisciplinary studies, incorporating art therapy, psychology, psychiatry, art, computer science, applied statistics, etc., be performed under the umbrella of a new field called Computational Art Therapy (CAT). The active participation of professionals in these fields is essential for the development of this type of system. Much time, effort, manpower, and cost will be required before its practical use can be envisaged however. Finally, the issue of whether human experts can be substituted by such a computer system should be briefly mentioned here. The proposed system can offer various types of information rapidly and accurately, and can help in the art therapists’ decision-making process. The final decision is up to the art therapists. It is not necessary for art therapists, however, to spend much time and effort to perform what a computer system can do for them.