eduzhai > Applied Sciences > Engineering >

Contact Area Detector using Cross View Projection Consistency for COVID-19 Projects

  • king
  • (0) Download
  • 20210507
  • Save

... pages left unread,continue reading

Document pages: 12 pages

Abstract: The ability to determine what parts of objects and surfaces people touch asthey go about their daily lives would be useful in understanding how theCOVID-19 virus spreads. To determine whether a person has touched an object orsurface using visual data, images, or videos, is a hard problem. Computervision 3D reconstruction approaches project objects and the human body from the2D image domain to 3D and perform 3D space intersection directly. However, thissolution would not meet the accuracy requirement in applications due toprojection error. Another standard approach is to train a neural network toinfer touch actions from the collected visual data. This strategy would requiresignificant amounts of training data to generalize over scale and viewpointvariations. A different approach to this problem is to identify whether aperson has touched a defined object. In this work, we show that the solution tothis problem can be straightforward. Specifically, we show that the contactbetween an object and a static surface can be identified by projecting theobject onto the static surface through two different viewpoints and analyzingtheir 2D intersection. The object contacts the surface when the projectedpoints are close to each other; we call this cross view projection consistency.Instead of doing 3D scene reconstruction or transfer learning from deepnetworks, a mapping from the surface in the two camera views to the surfacespace is the only requirement. For planar space, this mapping is the Homographytransformation. This simple method can be easily adapted to real-lifeapplications. In this paper, we apply our method to do office occupancydetection for studying the COVID-19 transmission pattern from an office desk ina meeting room using the contact information.

Please select stars to rate!


0 comments Sign in to leave a comment.

    Data loading, please wait...