/
Haptic identification of objects and their depictions Haptic identification of objects and their depictions

Haptic identification of objects and their depictions - PDF document

danika-pritchard
danika-pritchard . @danika-pritchard
Follow
392 views
Uploaded On 2017-04-06

Haptic identification of objects and their depictions - PPT Presentation

1 haptic perceptionnamely that the We thank Beall and Judith Loftus support from National BNS from the Centre the Ontario Information Research Centre Psychology UCSB Santa Barbara 1O6 ID: 336881

haptic perception-namely that the

Share:

Link:

Embed:

Download Presentation from below link

Download Pdf The PPT/PDF document "Haptic identification of objects and the..." is the property of its rightful owner. Permission is granted to download and print the materials on this web site for personal, non-commercial use only, and to display it on your personal computer provided you do not modify the materials and that you retain all copyright notices contained in the materials. By downloading content from our website, you accept the terms of this agreement.


Presentation Transcript

1, subjects, while haptically identified reduced cues compliance, mass, single outstretched finger, or five fingers. Performance decreased over superior to identification structure and integration performance was open-fingered gloves, information. Consequently, haptic perception-namely, that the We thank Beall and Judith Loftus support from National BNS- from the Centre the Ontario Information Research Centre Psychology, UCSB, Santa Barbara, 1O6-9660. of view" quite restricted objep a greater functional field contigu- ow fingerpads, depending on the condition. When the the one-finger aperture size mance that virtually identical that obtained touch-a mean and response la- Society, Inc. 172 KLATZKY, LOOMIS, LEDERMAN, WAKE, (Merman & Klatzky, Rosch et 1976; Tversky & from a category tend overall shape structure, so a photographic prototypical perspective is recognizable (Rosch et & were virtually the haptic object identification. the potential nonstructural cues, the had no parts, were table (hence mass could not any poten- (e.g., a detachable cover on a teapot) rigid. The subjects wore gloves also attenuated object's material. assess whether integration over fingertips occurred, ways: with a and with The subjects also attempted to name a set that were derived from photographs If it material information that underlies successful haptic to see haptic the level gloves and fixed of the stimuli greatly reduce information. The third possible explanation haptic greater structural haptic conditions will be superior to fiicture the fingers real objects, propo$ed the hand should produce superior performance co~npared with the use of a single there should extended fingers than with the contribution of information The subjects exl lored in the same ways as they did before, but now wi$their to the object's surface, so that Rroper- roughness and thermal kcarne available. To the extent that material is important in tification, differences exploratory conditions be reduced. contribution of increase as structure is creasingly limited. In a preliminary study, we simply assessed the performance available full-cue conditions: when the objects were explored touch (without an ungloved, dimensional raised pictures were viewed. The materials consisted a set 36 objects real and pictorial form. had no compliaryt parts were glued Picture stimuli used in Experiments derived from in the real- object conditions. 174 KLATZKY, LOOMIS, LEDERMAN, WAKE, AND FUJITA B.aellno Ra.1 Reel Real Basellno Plcture Plclure Real Free 5 1 Plctur. panel) and (right panel) the baseline conditions and the Experiments 1 Correlations were across objects Each object appeared in given exploratory two subjects. were averaged, scores were conditions over the set objects. The correlation Table 1. accuracy were expected. Intercorrelations were generally significant indicating common variance. The picture-5 some extent conditions. These correlations reflect com- at either perceptual levels (e.g., feature cod- (e.g., name-access time). find significant between-condition 1 condition does not seem lower variance floor effects reflect distinct pro- differed (with respect to or both) that information about is more haptically accessible the idea that curs across the fingers these stimuli. Although the present stimuli exploratory constraints severely lim- the object's material, real objects were still identified better the object enhanced the condition five fingers indicating integration real objects. pictures was fingers were used, relative to the baseline values pictures were was expected from previous research. However, less expected the finding that five were better than one picture identification. with the Loomis et al. (1991), who found no advantage in haptic fingers were comment further below. objects in real-5 and was within of mean the base- in which the subjects hand. How- neither condition showed mean RT the baseline sec; none of the sub- any real-object condition Correlations Across Accuracy in Experiment R1-RT R1-AC R5-RT R5-AC RF-RT RF-AC PI-RT PI-AC P5-RT RI-AC R5-RT R5-AC RF-RT RF-AC PI-RT P1-AC P5-RT P5-AC -.24 .23 -.37 .28 -.30 .24 .06 .06 -.58 Note-RT = response time, AC = accuracy, R1 = real-1, R5 = real-5. RF = real-free, P1 = picture-I, P5 = picture-5. Correlations of .35 and above are significant at .05. 176 KLATZKY, LOOMIS, LEDERMAN, WAKE, AND FUJITA and Accuracy Rl-RT Rl-AC RS-RT RS-AC RF-RT Rl-AC -.50 RS-RT .52 -. 13 R5-AC - .41 .13 - .73 RF-RT SO .20 .83 - .46 .32 - response time, R1 = real-I, R5 real-free. Cor- above are significant .05 level. 1, fingertip status uncovered, Experiment 2) (Although the subjects were randomly assigned two experiments from a the populations were very three effects were significant [F(2,40) = 24.60, p .0001], fingertip status [F(1,20) = 12.09, p .01], and for the interaction [F(2,40) = 4.88, p .05]. confirmed that the open glove and reduced of the two experiments were very sim- were virtually identical real-5 conditions. Hence, the corresponding analy- accuracy scores showed only [F(2,40) = 7.81, p .01]. The experiments allowed essentially is access to material. The relative difficulty the various would be to differ most, then, material can substantially contribute to recogni- in the particular, per- to differ between the two information is subjects explore single extended finger. This the means between the two experiments are further agreement for corresponding experiments.) These correlations indicate relatively difficult experiment were simi- larly difficult the less easily the subject can structural informa- tion about object. This Correlations between Experiments Respect to Measures on R5-RT RF-RT R I -AC R5-AC response time, R1 = real-1, R5 real-free. Correlations are significant. RTs, but, for for RT, the real- 1 should make information about the ob- iect's material available. due to 'kncoding its structural conducted a the two exueriments for conditions strultural was most limited; there- a substantial contribution periment 2 (but not in Experiment 1). A factor analysis performed on the basis of the Experiments 1 (which use the units of tion). Table 4 shows the solution orthogonal rotation, underlying factors the variance. The with high on a factor can some underlying determinant performance. suggested that the the one-finger exploration conditions in ~xperiments and that these condi- tions were distinct from the conditions high loadings five-finger conditions experiments, is likely represent the shape information, which would be highly those conditions. one-fqer that performance conditions differed from five fingers even Given the this factor reflect shape information, a different sort the five-finger representing one-finger perfor- in Experiment reliance on given impoverished access One might wonder whether the real-1 either experiment was correlated since these conditions the other sures. However, between real- l and pic- ture measures were generally low. This Experiment 1 The real- 1 accuracy measures Experiment 2 four picture periment 1 the eight r and then RT in 2 and 178 KLATZKY, LOOMIS, LEDERMAN, WAKE, real-object contours tive and promote integration sure, as tive cue. Relatively good performance a single outstretched finger attests the effectiveness this additional information. five outstretched fingers indicate that andlor additional advantage from indicates a contribution of kinesthetic simultaneously contacted also indications 2-D dis- contrast to our previous results (Loomis et al., 1991), slightly better a greater "aperture." This attributed to procedure: The aper- more extreme vs. five one vs. two fingers the previous study), subjects were required to fingers together. taneously sampled multiple-finger condition this study than it was also possible that a role exploration, rather than perceptual in- tegration in the a widened effective field achieved in the hands cues to terial were strongly reduced, indicate real objects can structural cues (shape with an The comparison tween the two in which the subjects fingertips either covered (Experiment provides further understanding of of an object's material its identification. The results suggest that when have access information about from their grasping, additional cutaneous informa- tion about material produces only a small facilitation. information is limited single exploring digit, the of ma- terial information has a more substantial effect bring the single- finger condition full hand. support the assumption that informa- tion about material contributes in the single-finger condition than in the other ex- ploratory conditions. factor analysis on correla- across objects indicated that there were three factors underlying the various real-object performance measures distinct factors for the the single-finger with and with- information available. have tentatively sug- gested that the represents shape infor- mation derived simultaneous contributions from the second factor represents con- tours extracted single finger and the third factor responds to the contribution certain about these attributions, factor analysis clearly indicates that the information identify an and multiple-finger differs not only when information about material is present but greatly attenuated. ijber taktilmotorische Figurwahrnehmung. Psy- chologische Forschungen, A theory of hu- image understanding. BIEDERMAN, I., a Ju, G. (1988). Surface versus edge-based determinants of visual recognition. Cognitive Psychology, 20, 38-64. CHAN, T.-C., CARELLO, C., a TURVEY, (1990). Perceiving ob- CRAIG, J. C. (1985). Attending to two fingers: & Psychophysics, 38, 496-511. DACEN, D., a COULSON, M. (1988). Tactile parative study. Tatham & A. G. Dodds (Eds.), Proceedings of rhe International Symposium on Maps Visually Handicapped People (pp. 7-23). M. A. (1989). pattern perception The advantage late blind. 18,379-389. for visual pat- a comparison with the Vlsion Re- search, 18, 1565-1571. KATZ, D. (1989). fie world of touch (L. E. Kmeger, Hillsdale, NJ: Erlbaum. KENNEDY, J. 8r Pictures to see Perkins & B. Leondar (Eds.), The arts and cognition (pp. 118-135). Baltimore: Johns Hopkins L., LEDERMAN, a "expert system." & Psycho- physics, 37, 299-302. LAPPIN, J. S., a FOULKE, Expanding the tactual field & Psychophysics, 14, 237-241. LEDERMAN, S. J., a KLATZKY, R. L. (1990). Haptic Knowledge-driven exploration. 8r (1990). Visual mediation and the haptic recognition of two- dimensional pictures of common objects. Perception & Psychophysics, 47, 54-64. LOOMIS, J., KLATZKY, R. L., a LEDERMAN, S. J. visual picture recognition LOOMIS, J., a Tactual perception. K. Boff, & J. Thomas (Eds.), of human per$ormance York: Wiley. MAGEE, (1980). Exploring pictures tactu- Rosc~, MERVIS, C., BOYES-BRAEM, P. (1976). Basic in natural categories. Cognitive Psychology, 8, 382-439. TVERSKY, B., & (1984). Objects, parts, Journal (1935). Tactile Tactile Japanese]. Japanese Journal of Psychology, 10, 327-390. (Manuscript received June revision accepted for January 13,