This page is a part of, a wiki devoted to computer vision. It focuses on low level computer vision, digital image analysis, and applications. It is designed as an online textbook but the exposition is informal. It geared towards software developers, especially beginners, and CS students. The wiki contains mathematics, algorithms, code examples, source code, compiled software, and some discussion. If you have any questions or suggestions, please contact me directly.

Computation error

From Computer Vision Primer

Jump to: navigation, search

Pixcavator computes numerous characteristics of objects (see Measuring objects). In terms of errors of computation they fall into four categories.



For most of the measurements Pixcavator uses only integer arithmetic, so there is no error. These are Area, Saliency, Contrast, Intensity (gray level). (Of course the resolution affects how well the digital image represents the reality and, therefore, the accuracy. See Continuous vs discrete.)

Rational numbers

Some measurements aren't integers but come as ratios, in fact pairs, of integers. This is almost as good because the number is turned decimal only when it is displayed. So the error can be easily managed. The location of the Centroid (or the Center of mass) is a ratio of Moments. The Average contrast is saliency divided by the Area. Average intensity also falls into this category.


Lengths of curves cannot be computed exactly in the digital environment. In Pixcavator this error is limited to 8% (more testing is needed). Therefore other characteristics also have this error: Perimeter, Roundness, Thickness and length.

Floating point

Pixcavator does no employ floating point arithmetic. For what can happen when you round up floating point numbers, see this article The Patriot Missile Failure.

Personal tools