Tactile classification of traditional and computerized media in three adolescents who are blind.
In the setting of this longitudinal study, we have been particularly interested in the activities of categorization of shapes and the attribution of meaning to shapes. Beyond the simple perception of shapes, we wanted to evaluate the capacity of these students to categorize geometric line drawings defined by several dimensions. In addition, we wanted to study and compare the categorization strategies used by these students when they performed identical tasks either with traditional material (on thermoformed paper) or with the Tactos system.
The Tactos platform (see Figure 1) consists of a graphics tablet and its pen, a computer, and a tactile stimulator (2 standard 8-pin electronic braille cells). The participant can explore a line drawing by moving the pen on the graphic tablet. The Tactos software provides the link between the position of the pen and the stimulators: The 16 (2 x 8) pins of the braille cells are activated in an all-or-none fashion whenever the participant encounters the shape (which is represented by black pixels) being explored by means of the pen. The tactile stimulation is delivered on the free hand. It is to be noted that perception with this instrument does not involve any symbolic coding--the activation of one or several pins on the braille cells corresponds solely to contact with the virtual object--and that at any one moment, access to the object (a 2-D shape in this case) is partial--the capture zone of 64 pixels is not large enough to cover the entire shape, so that global perception requires active displacements by the participant. After acquiring effective exploratory strategies (which takes about 20 hours of learning), the students were able to perceive and recognize simple shapes.
[FIGURE 1 OMITTED]
We conducted a systematic comparison of the categorization strategies that the three students used when exploring 2-D line drawings, which were presented either as raised forms on the traditional thermoformed paper or as computerized forms on the graphic tablet of the Tactos device. More precisely, the participants first performed a recognition exercise in each of these two contexts to verify that they could properly recognize the line drawings used for the categorization and that they were sensitive to the three dimensions that were used (size, orientation, and shape). Then, two categorization tasks were tested: a free-classification task and a sorting task into predefined categories. The main object of our study was to investigate the students' understanding of geometric line drawings and thus to compare the categorization in two contexts of perception: direct perception with bare fingers using the traditional material and instrumental perception using the Tactos device.
The participants were three high school students who were blind from birth and were aged 14 at the time of the observation. Two students (one boy, "Michael," and a girl, "Belinda") attended a community school, and the other girl "Janice" attended a specialized school for students who are visually impaired. Parental consent was obtained in writing prior to the implementation of the study.
The material to be recognized and categorized was composed of 2-D geometric line drawings (such as a square, rectangle, angle, triangle, and oval) that varied on two out of three dimensions at a time (shape and orientation, shape and size, or shape and orientation). The variation between two line drawings was 45 degrees for the orientation and 50% for the size. These line drawings were either printed on paper as raised forms (direct perception) or presented in computerized form that could can be read via the Tactos device (instrumental perception). For the initial recognition task, there were 14 different items. Each item consisted of 4 line drawings: the target line drawing was placed at the top of the sheet, and the 3 line drawings for comparison (an identical copy of the target form and two distracters) were placed underneath. For the free-classification task, there were 8 different items. Each item consisted of 4 line drawings: the reference drawing at the top and the 3 drawings for comparison underneath (see Figure 2). Two drawings were identical to the reference drawing on one dimension, but different on the other dimension (analytical choice); the third line drawing was globally similar to the reference drawing, but was not strictly identical to either of the dimensions (holistic choice).
[FIGURE 2 OMITTED]
For the task of sorting nine items into predefined categories, each item consisted of four line drawings placed in a square arrangement (see Figure 3). Each drawing varied in one or two dimensions relative to the others. Thus, in the example shown in Figure 3, line drawings 1 and 2 are identical in shape but not in size (the same is true for drawings 3 and 4), whereas line drawings 1 and 3 are identical in size but not in shape (the same is true for drawings 2 and 4).
The three participants were followed for about 20 hours, during which time they became familiar with the Tactos device and gained a certain level of ease and expertise in exploring the line drawings. They were able to identify complex shapes (such as a frieze) and to explore mathematical curves. After this learning period, we performed tasks of recognition, free classification, and sorting by category; all the tasks were performed using traditional paper (direct haptic perception) and the Tactos device (instrumental perception).
[FIGURE 3 OMITTED]
Concerning the 14 trials of the recognition task, the participants had to explore a target line drawing (such as a small square) and then find the identical line drawing among two additional distracters (for example, a medium-sized square and a large square). The responses could be correct or incorrect. The participants were not informed whether their responses were correct or erroneous.
For the eight trials of the free-classification task, the participants had to explore the target line drawing and then match it to one of the comparison line drawings (Cook & Odom, 1988). The choice could be analytical (the line drawings were identical on one dimension but different on the other) or holistic (the line drawings were globally similar, but not strictly identical on either dimension). In this task, there were no "correct" or "incorrect" responses; the result was necessarily either analytical or holistic.
For the categorical sorting task, on each of the nine trials the participants explored four line drawings (for instance, four triangles that varied in size and orientation) and had to sort them into two groups. If the participants sorted the line drawings according to just one of the two dimensions, the strategy was considered to be analytical; if they took both dimensions into account, the strategy was considered to be holistic. As in the previous classification task, there were no correct or incorrect responses; the result was necessarily either analytical or holistic. For each trial of each task, the participants were asked to explain their choice verbally.
The three participants correctly recognized 100% of the line drawings when using the traditional paper medium. In the computerized situation, the results were similar. Michael successfully recognized 100% of the line drawings using Tactos. Janice made just one error when faced with squares of various sizes (93% correct responses). Belinda committed two errors with curved line drawings (86% correct responses). Overall, it seems fair to conclude that recognition is as efficient in the instrumental situation as with traditional direct haptic perception.
Overall, the majority of strategies were analytical, whatever the task or the mode of perception (see Table 1). The condition that produced the highest frequency of holistic strategies was the categorical sorting with Tactos, but it should be emphasized that even in this case, the proportion remained low (15%).
For the individual results of each participant, we observed some differences. Michael, for example, exclusively used an analytical strategy whatever the task and whatever the mode of perception (see Table 2). Similarly, Belinda exclusively used the analytical strategy in direct perception, but made one holistic choice in each task with Tactos. Janice used the holistic strategy once for free classification on traditional paper and three times for categorical sorting with Tactos.
With regard to the dimension used in the case of analytical choices, Michael and Janice always used the dimension "shape" whenever it was involved (whether associated with size or orientation). When size and orientation were the two dimensions involved, their choice was based on orientation. Belinda based her analytical choices on shape when it was associated with orientation, but not when it was associated with size. When the material was defined by the dimensions' size and orientation, her choices were distributed equally between the two.
The results of the study show that the Tactos device is appropriate for reading and recognizing geometric line drawings by adolescents who are blind. In all, only three mistakes were made: one concerning a change in size and two concerning a change in shape. Even if instrumental perception seems to be more difficult than direct perception (for which there were no mistakes), we found again that the performances already observed with Tactos were good (Rovira & Gapenne, 2008).
With regard to the categorization tasks, the results also show that the three participants preferred an analytical mode of classification in each of the two exploratory contexts (traditional and with Tactos). This finding is in agreement with the results usually obtained with adults in a visual situation (Berger, 1992) or in a tactile situation (Berger & Hatwell, 1996; Schwarzer, Kufer, & Wilkening, 1999). It is true that the situation strongly favored an analytical strategy, because our material varied only on two dimensions and there was no time limit (Berger, 1992). In this context, the rare holistic responses may have been a sign that the participants had some difficulty understanding the line drawings: Nevertheless, the important point here is that the use of Tactos does not seem to lead to a radically different style of task management.
The results thus provide evidence that the appropriation and use of Tactos improve the possibility of learning to read two-dimensional graphic objects, particularly in computerized environments. In line with this point, the participants were able to use potentially all the graphical tools to produce a digital line drawing and to explore it immediately by means of Tactos in order to perceive and understand its properties. Other substitution devices, such as the Optacon (Bliss, Katcher, Rogers, & Shepard, 1970), which was designed to make it possible to read written characters, exist. Thus, Tactos may be considered a complementary system, which makes it possible to explore pictures that accompany written text. The pedagogical interest of having recourse to images in certain learning contexts is undeniable. In addition, Tactos was designed for reading computerized objects, which is not the case for the Optacon.
Finally, as this study has shown, Tactos does not limit the diversity of students' activities, and so respects the individual modes in making sense of graphical objects. However, to confirm these preliminary results, we plan to implement a systematic study with a larger sample.
We express our warm thanks to all the staff of Normandie Lorraine (76 240 Le Mesnil Esnard) and to the students for their collaboration. We also warmly thank John Stewart for his translation of the text and his comments.
Bach-y-Rita, P. (1972). Brain mechanisms in sensory substitution. New York: Academic Press.
Berger, C. (1992). Perception analytique et globale [Analytical and holistic perception]. L'Annee Psychologique, 92, 105-136.
Berger, C., & Hatwell, Y. (1996). Developmental trends in haptics and visual free classifications: Influence of stimulus structure and exploration and decisional processes. Journal of Experimental Child Psychology, 63, 447-465.
Bliss, J. C., Katcher, M. H., Rogers, C. H., & Shepard, R. P. (1970). Optical-to-tactile image conversion for the blind. IEEE Transactions on Man-Machine Systems, 11(1), 58-65.
Cook, G. L., & Odom, R. D. (1988). Perceptual sensitivity to dimensional and similarity relations in free and rule-governed classification. Journal of Experimental Child Psychology, 45, 319-338.
Gapenne, O., Rovira, K., Ali Ammar, A., & Lenay, C. (2003). Tactos: A special computer interface for the reading and writing of 2-D forms in blind people. In C. Stephanidis (Ed.), Universal access in HCI: Inclusive design in the information society (pp. 1270-1274). London: Lawrence Erlbaum.
Lenay, C., Gapenne, O., Hanneton, S., Marque, C., & Genouelle, C. (2003). Sensory substitution, limits, and perspectives. In Y. Hatwell, A. Streri, & E. Gentaz (Eds.), Touch for knowing (pp. 275-292). Amsterdam: John Benjamins.
Rovira, K., & Gapenne, O. (2008). Categoriser a l'aide d'un systeme de suppleance perceptive: Strategies d'exploration et de classification d'adolescents aveugles [Categorizing with the help of a sensory substitution device: Exploration and classification strategies in blind teenagers]. In E. Loarer, P. Vrignaud, J. L. Mogenet, F. Cuisinier, H. Gottesdiener, & P. Mallet (Eds.), Perspectives differentielles en psychologie (pp. 399-402). Rennes, France: Presses Universitaire de Rennes.
Schwarzer, G., Kufer, I., & Wilkening, F. (1999). Learning categories by touch: On the development of holistic and analytic processing. Memory and Cognition, 27, 868-877.
Sribunruangrit, N., Marque, C., Lenay, C., Gapenne, O., & Vanhoutte, C. (2004). Speed-accuracy trade-off during performance of a tracking task without visual feedback. IEEE Transactions on Neural Systems and Rehabilitation Engineering, 12, 131-139.
Ziat, M., Gapenne, O., Stewart, J., & Lenay, C. (2007). Haptic recognition of shapes at different scales: A comparison of two methods of interaction. Interacting with Computers, 19, 121-132.
Katia Rovira, Ph.D., professor, Department of Psychology, University of Rouen, 76821 Mt. St. Aignan, Cedex, France; e-mail: <katia.rovira@ univ-rouen.fr>. Olivier Gapenne, Ph.D., professor, Department of Technology and Human Sciences, University of Technology of Compiegne, Centre Pierre Guillaumat, BP 319, 60206 Compiegne Cedex, France; e-mail: <olivier.gapenne@ utc.fr>.
Table 1 Mean percentages of analytical choices in the two tasks and the two perceptual situations. Analytical Free Categorical choice classification sorting Total Haptic 96 100 98 Tactos 96 85 91 Total 96 93 93 Table 2 Percentage of analytical strategies as a function of tasks and perceptual modes for each participant. Analytical strategy Michael Belinda Janice Haptic Free classification 100 100 88 Categorical sorting 100 100 100 Tactos Free classification 100 88 100 Categorical sorting 100 89 67 Total 100 94 89
|Printer friendly Cite/link Email Feedback|
|Title Annotation:||Research Reports|
|Author:||Rovira, Katia; Gapenne, Olivier|
|Publication:||Journal of Visual Impairment & Blindness|
|Date:||Jul 1, 2009|
|Previous Article:||The development of a universal tangible symbol system.|
|Next Article:||"Impact of Curb Ramps on the Safety of Persons Who Are Blind.".|