A Comparison of Hardware Implementations for Low-Level Vision Algorithms
Abstract
Early and intermediate vision algorithms, such as smoothing and discontinuity detection, are often implemented on general-purpose serial, and, more recently, parallel computers. The excessive time required by these general- purpose computers prevents real-time computation of these vision algorithms. Special-purpose hardware implementations of low-level vision algorithms may be needed to achieve real-time processing. This memo reviews and analyzes some hardware implementations of low-level vision algorithms. Two types of hardware implementations are considered: the digital signal processing chips of Ruetz (and Broderson) and the analog VLSI circuits of Carver Mead. Both these approaches claim to achieve real-time image processing; both have limited the vision problem that they solved in ways largely inconsistent with vision processing in unrestricted environments. The advantages and disadvantages of these two approaches for producing a general, real-time vision system are considered. As early attempts at comprehensive vision hardware, these two approaches provide useful insights for future developments of vision hardware.
Document Details
- Document Type
- Technical Report
- Publication Date
- Nov 01, 1989
- Accession Number
- ADA223699
Entities
People
- Ed Gamble
Organizations
- Massachusetts Institute of Technology