Vision driven Biomechanical Human Control

shapeimage_1
 
Biomechanical Human Control

 
We did following 2 projects. Main theme was to implement vision function onto the biomechanical human model.

1. Real and Virtual Face interaction
Biomechanical Neck Model control based on the captured real human face location by web cam.

.2.Biomechanical Human Body Control driven by the vision information in the virtual environment.

The below is the demo for each project.

1. Virtual and Real Face Synch
* We use OpenCV to detect the real human face using the captured image with web cam. This algorithm uses HOG feature for the face detection, and return the detection square. Based on the location of the real face, virtual face try to gaze and move the neck with neural controller.

 
* Virtual Character detect 2 faces and gaze at the one closer
 
 


2. Vision Driven Arm Control with IK

* Tow images at the beginning of the video 2 are the image of the target object captured by the eye (camera set on eye). Left side is the original captured scene, and the right side is the detection result. Red square shows that the algorithm find the one as a target. We use OpenCV for the contour detection, and IK for the control of the arm.

*
Project Report
*Project Presentation