Automatic Estimation of Multidimensional Ratings from a Single Sound-Symbolic Word and Word-Based Visualization of Tactile Perceptual Space

2019

ISASE

Self Cite

Section: Discussionmentioning

confidence: 99%

Image Retrieval System Using Japanese Sound Symbolic Words for Surface Texture

Yamagata

Kagitani

2019

ISASE

Self Cite

“…Furthermore, tactile perceptual dimensions have been studied in a large body of psychophysics literature (e.g., Yoshida, 1968 ; Hollins et al, 1993 ; Gescheider et al, 2005 ; Bergmann Tiest and Kappers, 2006 ; Chen et al, 2009 ) Most psychophysical studies on the perceptual space of touch have used the multidimensional scaling method and/or the semantic differential method. However, analyzing the sensory vocabulary of touch could be an alternative and effective way of investigating categories of tactile perception (see also Guest et al, 2011 ; Dingemanse and Majid, 2012 ; Doizaki et al, 2017 ). In this study, we investigated associations between syllables of Japanese sound-symbolic expressions and fundamental tactile perceptual qualities.…”

Section: Discussionmentioning

confidence: 99%

Bouba/Kiki in Touch: Associations Between Tactile Perceptual Qualities and Japanese Phonemes

Watanabe

2018

Front. Psychol.

Self Cite

Several studies have shown cross-modal associations between sounds and vision or gustation by asking participants to match pre-defined sound-symbolic words (SSWs), such as “bouba” or “kiki,” with visual or gustatory materials. Here, we conducted an explorative study on cross-modal associations of tactile sensations using spontaneous production of Japanese SSWs and semantic ratings. The Japanese language was selected, because it has a large number of SSWs that can represent a wide range of tactile perceptual spaces with fine resolution, and it shows strong associations between sound and touch. In the experiment, we used 120 everyday materials to cover basic material categories that could be associated with fundamental dimensions of tactile perception. Upon contact with these materials, participants expressed their tactile sensations by using Japanese SSWs, and at the same time, evaluated the tactile sensations by semantic differential scales using adjective pairs. Thanks to the variety of testing materials, we were able to demonstrate the existence of systematic associations between sounds and tactile fundamental perceptual dimensions in a more detailed and comprehensive way than ever done so before. In particular, we found that for vowels, positive tactile ratings were associated with the back vowel (/u/), while negative ratings were associated with the front vowels (/i/ and /e/). The central vowels (/o/ and /a/) were mainly associated with rough, hard, and dry feelings. Consonants were categorized based on vocal features and articulation. The category of the voiced consonants (e.g., /dz/ and /g/) corresponded to feelings of roughness, while that of voiceless consonants (e.g., /ʦ/, and /s/) corresponded to feelings of smoothness. The categories of the bilabial plosive (/p/ and /b/) and voiced alveolar nasal (/n/) consonants were mainly related to soft, sticky and wet feelings, while that of voiceless alveolar affricate (/ʦ/) and voiceless velar plosive (/k/) consonants were related to hard, slippery and dry feelings.

“…Many studies have used this method to analyze TPS. Our previous study (Doizaki et al, 2017) differed from previous work in that we proposed a method for estimating evaluations of touch using only a single word. Specifically, we constructed a system that can convert a word that intuitively expresses tactile sensations into an information equivalent derived from evaluations of 26 pairs of touch adjectives.…”

Section: Introductionmentioning

confidence: 95%

Visualizing Individual Perceptual Differences Using Intuitive Word-Based Input

Watanabe

2019

Front. Psychol.

Self Cite

Numerous studies have investigated the fundamental dimensions of human tactile perceptual space using a wide range of materials. Participants generally touch materials and quantitatively evaluate variations in tactile sensations for pairs of adjectives pertaining to the material properties, such as smooth—rough and soft—hard. Thus, observers evaluate their perceptual experiences one by one in terms of adjective pairs. We previously proposed an alternative method of qualitative evaluation of tactile sensations. Our system can automatically estimate ratings of fundamental tactile properties from single sound-symbolic words. We were able to construct a word-based perceptual space by collecting words that express tactile sensations and applying them to the system. However, to explore individual differences in perceptual spaces, different databases for converting words into ratings of adjective pairs are required for each individual. To address this, in the present study we created an application that can automatically generate an individualized perceptual space by moving only a few words in the initial word-based perceptual space. In addition, we evaluated the efficacy of the application by comparing the tactile perceptual space before and after use.