|
Computer Vision and Image Analyses applications require an extensive and diverse
range of skills. PantherVision is fortunate in that we have extensive practical
experience of all key areas and this means that we have the capability to address
any computer vision problem. The following outlines just some of the areas where
we have a unique level of experience. Where possible, links are included to more
detailed descriptions.
Low Level Image Analyses covers all the basic techniques that might be needed to
underpin any computer vision applications, for example, edge detection; region segmentation;
thresholding; pixel based operations; colour or grey level analyses; histogram analyses
and many more.
Medium Level Image Analyses includes techniques such as pixel based data fusion;
feature extraction and classification; feature merging;
High Level Image Analyses takes the results of more basic processing and inteprets
the results and is usually one of the final stages in any application. For example,
extracted features may be grouped and interpreted as belonging to a particular object
such as a car, and the speed of the car inferred from the way in which the features
move over time. Adaptive and dynamic feedback enables techniques to tune their parameters
to optimise their own performance.
The analyses of video sequences requires other levels of expertise including, for
example, the ability to track objects over time through the sequence. Special attention
often needs to be given to the implementation to ensure that it operates at an acceptable
speed. For some applications this means in real-time and generally requires careful
thought throughout the development.
Any particular component may be used at different levels in different applications
or at different levels of complexity.
|