If we have a number of different objects, human can tell the the biggest/largest object within no time. How does human eye compare object sizes? And why does it happen so fast. On the other hand, a computer could take "some time" to tell the largest object apart? So i suppose human eyes beat the computer. Are there any mathematical computations that can assists why this happens fast in human and can we borrow this apply it in computer vision?
This deals with Stereo vision of human eyes. Please see links below:
https://en.wikipedia.org/wiki/Computer_stereo_vision
http://www.vision3d.com/vocvis.html#depth
Still better text on stereo vision can be consulted in:
Milan Sonka, Vaclav Hlavac, and Roger Boyle. Image processing, analysis, and machine vision. Cengage Learning, 2014.
Thanks Kane. That was so helping. I have seen that stereo vision is very important in determining size and distance of objects. If we had a number of objects, would we enable the computer to do it as fast the human eyes? I read the attached paper and got interested.
Thanks.
Dear Dr. Kengo
Typical stereo vision arrangements make use of two monocular cameras placed side by side and aligned with each other. The algorithms used for object tagging and comparison are quite slow.
A solution in this context would be in the recently available depth sensors like Microsoft Kinect and Asus ActionPro. I am working with Kinect. Though I am not focusing the object comparisons for their sizes directly, work is similar. This would definitely help you.
Hi, you can also find more about stereo vision and how it works in details here:
https://www.researchgate.net/publication/286778168_A_Study_in_3D_Structure_Detection_Implementing_Forward_Camera_Motion
Hope it will help you.
Thesis A Study in 3D Structure Detection Implementing Forward Camera Motion
- Badges
- Science topic
- Similar topics
- Linguistics
- Composition Studies
- Publications