Datasets
Standard Dataset
Experimental Data: Finger Scanning Speed and Heart Rate Counting
- Citation Author(s):
-
NorihisaMiki
Keio University - Submitted by:
- Norihisa Miki
- Last updated:
- Mon, 02/26/2024 - 02:18
- DOI:
- 10.21227/h8tj-q992
- Data Format:
- License:
94 Views- Categories:
- Keywords:
Abstract
Finger Scanning Experiments: Participants conducted experiments while seated and wearing an eye mask to eliminate visual information. They were instructed to horizontally scan the surface back-and-forth eight times with one of their fingers to assess surface roughness. The finger's motion was optically captured at a frame rate of 60 fps. Scanning speed was determined by measuring finger positions at each frame and dividing them by the frame length. Image analysis was performed using OpenCV, where the finger outline was extracted from the video. The analysis comprised three steps: 1) Grayscale conversion of the video, 2) Binarization of the video (extracting only skin-colored parts), and 3) Utilizing the center of the x-coordinate and y-coordinate of the contour extracted in step 2 as the finger's center (indicated by the star in Fig. 3), tracking the center point of the finger. Throughout the 8 back-and-forth motions, we anticipated that finger scanning motion would converge toward a pattern associated with the surface features. We investigated the average finger speed during the 8th scan and examined the change in scanning speed across the repeated back-and-forth motions. Participants randomly traced eight tactile samples, and three trials were conducted for each tactile sample. This resulted in a total of 24 trials for each participant.
Heart Rate Counting: Participants, while seated and wearing an eye mask, were instructed to count the number of their heartbeats without relying on external cues and verbally report it. The task was conducted in three intervals for each participant: 15, 30, and 45 s. The correct number of heartbeats was recorded using a wearable device (Fitbit Inspire2, Fitbit Inc., Delaware, US). The difference between the reported number and the correct number was divided by the correct heart rate, and the result was subtracted from 1, providing a metric to assess the accuracy of interoception. A more accurate interoception yields a value closer to 1.
Finger Scanning Experiments: Participants conducted experiments while seated and wearing an eye mask to eliminate visual information. They were instructed to horizontally scan the surface back-and-forth eight times with one of their fingers to assess surface roughness. The finger's motion was optically captured at a frame rate of 60 fps. Scanning speed was determined by measuring finger positions at each frame and dividing them by the frame length. Image analysis was performed using OpenCV, where the finger outline was extracted from the video. The analysis comprised three steps: 1) Grayscale conversion of the video, 2) Binarization of the video (extracting only skin-colored parts), and 3) Utilizing the center of the x-coordinate and y-coordinate of the contour extracted in step 2 as the finger's center (indicated by the star in Fig. 3), tracking the center point of the finger. Throughout the 8 back-and-forth motions, we anticipated that finger scanning motion would converge toward a pattern associated with the surface features. We investigated the average finger speed during the 8th scan and examined the change in scanning speed across the repeated back-and-forth motions. Participants randomly traced eight tactile samples, and three trials were conducted for each tactile sample. This resulted in a total of 24 trials for each participant.
Heart Rate Counting: Participants, while seated and wearing an eye mask, were instructed to count the number of their heartbeats without relying on external cues and verbally report it. The task was conducted in three intervals for each participant: 15, 30, and 45 s. The correct number of heartbeats was recorded using a wearable device (Fitbit Inspire2, Fitbit Inc., Delaware, US). The difference between the reported number and the correct number was divided by the correct heart rate, and the result was subtracted from 1, providing a metric to assess the accuracy of interoception. A more accurate interoception yields a value closer to 1.
