お知らせ

information

Ubiquitous Virtual Reality Lab

2025年7月10日

アルスホーム

KAIST virtual roughness study

This paper presents a visual inspection method that can eliminate the influence of self-affine grinding surfaces. Accuracy and Loss curves for each model using different datasets (A) Accuracy curve for D1 (B) Loss curve for D1 (C) Accuracy curve for D2 (D) Loss curve for D2 (E) Accuracy curve for D3 (F) Loss curve for D3. In these four correlation functions, the average brightness of image pixels increases with an increase in the roughness parameter. Correlation functions for the G1 interval (A), the G2 interval (B), the G3 interval (C) and the G4 interval (D). The Figure 12 depicts the fitting curves of roughness and brightness for each classification Computer vision interval. Processing equipment and measuring equipment (A) ordinary horizontal axis moment table surface grinder M7150H (B) surface roughness tester TA260.

In engineering, two-scale fractal models were widely used for treating with approximate similarity 44–47. However, this similarity is approximately valid for a certain range of Facial recognition system scales. The fractal nature of the grinding surface topography refers to the fact that the grinding surface maintains an intrinsically similar structure at different observation scales.

Sci-Tech NOW KAIST develops VR collaboration tech that visually shares touch – DongA Science

Sci-Tech NOW KAIST develops VR collaboration tech that visually shares touch.

Posted: Tue, 03 Nov 2026 08:00:00 GMT source

The Ubiquitous VR research aims at the development of new computing paradigms for Holistic DigiLog Life in Smart Space. Computing technology is radically changing the manner in which we communicate with media, spaces, people as well as computers. Behind every published article stand expert reviewers safeguarding science. Any product that may be evaluated in this article or claim that may be made by its manufacturer is not guaranteed or endorsed by the publisher.

  • The team previously received the Best Paper Award in 2024 for research on mixed reality (MR) remote collaboration.
  • What if you could see how rough an object feels, even without touching it?
  • Participants without haptic devices also reported greater confidence in their judgments and a stronger sense of contribution when the cues appeared on their own virtual hands.
  • In 2025, two of its papers—one on spatial sharing and another on visual and tactile experiences—received Best Paper Awards.

3 Correlation between image pixel point brightness and surface roughness

This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). Roughness, fractal-like property, grinding surface, fractal imaging, deep learning, two-scale fractal theory Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher. All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the https://aicurated.com/ editors and the reviewers. The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest. Concurrently, researchers may endeavor to integrate binocular reconstruction technology with the illumination method proposed in this paper, thereby investigating the broader potential applications of this fusion technology in the domain of surface reconstruction.

Turning roughness into visual cues

Combining the use of the three tools, several behaviors, patterns, and analyses of the sampled surfaces can be inferred. The roughness signature https://aicurated.com/computer-vision-and-perception/kaist-makes-virtual-roughness-visible-for-shared-vr-tasks/ also allows an overall surface evaluation, comparing one surface with the other surfaces of the same block and with all surfaces sampled at the same firing temperature. The other information (histogram and roughness graph) are also used to aid in the analysis of the coefficient behavior.

Visual Computing

Nevertheless, only one scale will result in a reduction in the accuracy of the deep learning model in predicting the roughness of the grinding surface due to the effect of the fractal nature of the grinding surface topography. Grinding surface images at the micrometer scale can increase the accuracy and enhance the stability of the texture features observed in the images. Sorry, a shareable link is not currently available for this article. If material is not included in the article’s Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. The evaluation of the surface quality of materials through the measurement and analysis of roughness parameters is known6,11,15 to be an effective way of determining the quality or standardization of surfaces.

  • Beyond speed and accuracy, the study also explored the psychological impact of the visual cues on the participants.
  • Rather than attempting to reproduce the actual tactile sensation on the skin, the system uses the outline’s shape and movement as visual cues.
  • The findings suggest that visualizing tactile information can help users make shared decisions even when only one partner has access to haptic feedback, effectively leveling the playing field in collaborative tasks.
  • Pairs completed the sorting task approximately 23% faster when roughness cues were displayed on both participants’ virtual hands compared to the condition without visual cues.
  • Users without haptic devices reported greater confidence in their judgments and a stronger sense of contribution when the cues appeared on their own virtual hands.

However, due to the change in surface roughness, the diffuse reflection coefficient and specular reflection coefficient of the grinding surface of the same material will change, which will further affect the image pixel brightness. The direct application of the aforementioned derivation to the calculation of the height of each point on the surface will inevitably result in the generation of erroneous data. Furthermore, the relative position relationship between the surface crest height can be calculated by the pixel brightness, which provides a foundation for the subsequent detection method. The ellipses fitted by this fitting method are typically represented in the form of standard ellipse equations. The angle between the surface normal of the ellipse and the Y-axis is , which can be expressed as Eq. Here, represents the incident light intensity, represents the diffuse reflection coefficient, and represents the Angle between the incident light and the surface normal.

KAIST virtual roughness study

KAIST virtual roughness study

In order to accurately simulate the change of the reflected light intensity with the roughness of the grinding surface, the Lambert illumination model is selected to construct the surface reflection model in this paper. Subsequently, the point cloud was clustered using the K-means algorithm in order to ascertain the height of each wave peak. This will be followed by a mathematical derivation to prove that there is a certain relationship between surface height and image brightness. Figure 1 illustrates that, at the macroscopic scale, images with different roughness levels are susceptible to light source noise, yet there are still discernible differences between them. Images acquired at macroscopic and microscopic scales using conventional imaging methods, respectively.

KAIST virtual roughness study

KAIST virtual roughness study

These include the grinding wheel particle size number, grinding wheel speed, grinding depth, axial feed parameters and transverse feed speed. The sample under examination was produced using the M7130 (as shown in Figure 10A) precision horizontal axis grinder, and the workpiece material is 45# steel, a Robotic process automation common engineering material. A comparison of the calculated value with the measured value has demonstrated the rationality and accuracy of the method proposed in this paper. Once the neural network has predicted the roughness level of an image with unknown roughness, this paper inputs the level into the corresponding index function in order to calculate the specific roughness value.