Haptic saliency model for rigid textured surfaces
Journal: Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume: 10893 LNCS
When touching an object, we focus more on some of its parts rather than touching the whole object’s surface, i.e. some parts are more salient than others. Here we investigated how different physical properties of rigid, plastic, relieved textures determine haptic exploratory behavior. We produced haptic stimuli whose textures were locally defined by random distributions of four independent features: amplitude, spatial frequency, orientation and isotropy. Participants explored two stimuli one after the other and in order to promote exploration we asked them to judge their similarity. We used a linear regression model to relate the features and their gradients to the exploratory behavior (spatial distribution of touch duration). The model predicts human behavior significantly better than chance, suggesting that exploratory movements are to some extent driven by the low level features we investigated. Remarkably, the contribution of each predictor changed as a function of the spatial scale in which it was defined, showing that haptic exploration preferences are spatially tuned, i.e. specific features are most salient at different spatial scales.
Haptic Saliency Model for Rigid Textured Surfaces
Journal: HAPTICS: SCIENCE, TECHNOLOGY, AND APPLICATIONS, PT I
Source: Web of Science (Lite)