Posts

Showing posts with the label parameterization

Hard hard user interfaces.

Started as someone doing ActionScript in Macromedia Flash I got into Human-Computer-Interaction after watching Jeff's Hann's multi-touch work. That's the power of a really cool demo that inspired the birth of NUIGroup community which spawned 1000s of makers all around the world making their own projector camera multitouch systems. All of it became mainstream once Apple released the iPhone to the world, and also gave birth to my interest in human-centered engineering. I spent some 10 years after that playing with sensors, haptics, and gestural interfaces through multiple input modalities but nothing stuck as much as multi-touch did. Thanks to Apple's execution. Everyone's phone/tablet interaction is default multitouch. (unless you have a visual impairment). Primsense evolved into Kinect, then came Wiimote, Leap but nothing stuck.  for speech, it was Amazon's Alexa. Latest news , that its going to lose Amazon 10B.  Like I am writing this post through my keyboard a...

CV dilemma

Looks like there's a problem with OpenCV's distributed parameterization of coordinates used during the camera calibration tasks. Three distinct sources of information on img distortion formulae apparently give three non-equivalent description of the parameters and equations involved: (1) In their book "Learning OpenCV" write regarding lens distortion: xcorrected = x * ( 1 + k1 * r^2 + k2 * r^4 + k3 * r^6 ) + [ 2 * p1 * x * y + p2 * ( r^2 + 2 * x^2 ) ], ycorrected = y * ( 1 + k1 * r^2 + k2 * r^4 + k3 * r^6 ) + [ p1 * ( r^2 + 2 * y^2 ) + 2 * p2 * x * y ], where r = sqrt( x^2 + y^2 ). Assumably, (x, y) are the pixel coordinates (in pixel units) in the uncorrected captured image corresponding to world-point objects with coordinates (X, Y, Z), camera-frame referenced, for which xcorrected = fx * ( X / Z ) + cx and ycorrected = fy * ( Y / Z ) + cy, where fx, fy, cx, and cy, are the capturing camera's intrinsic parameters. Therefore, having (x, y) from a ...