TactDeform: Finger Pad Deformation Inspired Spatial Tactile Feedback for Virtual Geometry Exploration

TL;DR

TactDeform introduces a dual-context electro-tactile rendering approach that translates interaction context (Approaching, Contact, Sliding) and geometric context (Face, Edge, Corner, Texture) into spatio-temporal fingertip patterns. Using a 32-electrode finger-worn array, the system achieves high accuracy in geometric feature identification ($85.7\%$) and texture discrimination ($95.8\%$) in a two-phase user study with $N=24$, without mechanical force feedback. Key contributions include the deformation-based encoding grounded in mechanoreceptor-inspired mappings, velocity-dependent pattern generation, and an open-source implementation to support VR geometry exploration and potential accessibility applications. The results suggest that parameterized, context-aware electro-tactile feedback can convey rich geometric and textural information, enabling naturalistic exploration across simple to complex 3D shapes. These findings advance tactile rendering beyond uniform or single-context schemes and point to adaptive, multi-modal, and inclusive XR haptics future directions.

Abstract

Spatial tactile feedback can enhance the realism of geometry exploration in virtual reality applications. Current vibrotactile approaches often face challenges with the spatial and temporal resolution needed to render different 3D geometries. Inspired by the natural deformation of finger pads when exploring 3D objects and surfaces, we propose TactDeform, a parametric approach to render spatio-temporal tactile patterns using a finger-worn electro-tactile interface. The system dynamically renders electro-tactile patterns based on both interaction contexts (approaching, contact, and sliding) and geometric contexts (geometric features and textures), emulating deformations that occur during real-world touch exploration. Results from a user study \rr{(N=24)} show that the proposed approach enabled high texture discrimination and geometric feature identification compared to a baseline. Informed by results from a free 3D-geometry exploration phase, we provide insights that can inform future tactile interface designs.

Figures (5)

• Figure 1: TactDeform Concept. TactDeform leverages a dual-context approach that combines interaction contexts (left) and geometric contexts (right) to generate parametric spatio-temporal tactile patterns. Interaction contexts include approaching, stationary contact, and sliding across surfaces, with associated parameters of velocity, orientation, and direction. Geometric contexts comprise feature types (faces, edges, corners) and texture levels (smooth, rough, rougher). This approach dynamically combines these contexts to render appropriate electro-tactile patterns that emulate natural finger pad deformations during virtual object exploration.
• Figure 2: Finger pad deformation visualization during different interaction contexts. (A) Finger pad approaches the tip of a teapot (a corner). (B) Finger pad approaches the body of the teapot (a face). (C) Finger pad approaches the ridge between the body and lid of the teapot (an edge). (D) Moving across the lid of the teapot, showing progressive deformation patterns across the contact area. (E) Finger pad remains stationary on the lid of the teapot, biased slightly to the left, with finger pad deformation reflecting this angular bias.
• Figure 3: The system integrates four components: (A) Desktop PC running Unity application connected via Quest Link to (B) Meta Quest 3 HMD for hand tracking and visual rendering, (C) Control circuit in a 3D-printed enclosure communicating via serial connection, (D) Flexible PCB tactile interface attached to the index finger pad with 32 electrodes in a 6$\times$6 configuration, and (D1) Enlarged view showing electrode specifications with 1.4mm diameter and 2.0mm center-to-center spacing. The flexible PCB extends from the palm side for minimal movement interference. Medical tape for finger attachment not shown for clarity.
• Figure 4: User Study Design Overview. (A) Overview of the two-phase study' procedure, validating TactDeform's concept through controlled tasks (Phase 1) and free exploration (Phase 2). (B) Electrode activation patterns showing how TactDeform renders different geometric features through spatial stimulation on the 32-electrode array. (C) A participant experiencing TactDeform during the study, wearing the Meta Quest 3 HMD with the finger-worn tactile interface attached to their index finger.
• Figure 5: Phase 1 Results. (1) Confusion matrix for Task 1 geometric feature classification. Participants classified haptic stimuli into Face, Edge, or Corner categories. Values show the percentage of trials where the true stimulus (row) was classified as the selected category (column). Overall accuracy was 0.857 across all participants. (2) Mean angle range explored by preference groups in Task 2. Participants were categorized into Orientation-preference and Uniform-preference groups based on their preferred patterns. Bars show mean angle range (in degrees) with standard error for each group-object combination. Blue bars represent the Orientation group, pink bars represent the Uniform group. (3) Pairwise preference matrix for Task 3 texture discrimination. The matrix shows the probability that the row stimulus was chosen as rougher than the column stimulus. Stimuli are ordered by increasing roughness: Lv1 Smooth, Lv2 Rough, Lv3 Rougher. Values represent percentages with overall accuracy of 0.958.