Computer vision and its application to APL

  • Authors:
  • I. Feldberg

  • Affiliations:
  • John Hopkin Univ., Laurel, MD

  • Venue:
  • APL '88 Proceedings of the international conference on APL
  • Year:
  • 1987

Quantified Score

Hi-index 0.00

Visualization

Abstract

Computer Vision refers to the application of human vision techniques to a computer, simply put, teaching the computer to see. This branch of computer science is well rooted in several other branches, including: image processing, artificial intelligence, and computational geometry. From these fields, as well as its own, come many interesting algorithms that can be used to form computer vision systems. These systems have been used in the past to perform such operations as inspecting tools on an assembly line and maneuvering a vehicle without human intervention. These tasks require two very different types of computer vision systems. In the first, a very limited domain can be covered, that is, the computer has to recognize only a few things (e.g. tool, belt, background). In the second, the computer has to be much smarter. It must recognize many diverse things (ground, walls, sky, obstacles).In this project, a medium-scale computer vision task was chosen. We were to take an image of a mandrill (the famous one) and create algorithms to accurately label parts of the image. After some thought, APL was chosen as the language to program the task in. Not only was APL easy to work with, but it proved to be a creative factor. In APL, we were able to use array techniques that were simply unavailable in any other language.