1 | initial version |
Well, SVM is a machine learning technique that is not specific to vision applications. Having said that, SVM may well be used for DSP applications (and plenty of others). The implementation in OpenCV is not specific to images, but requires the data being formatted as Mat
structures, which makes possible the use of the algorithm for other applications too. In other words, your data (are they features extracted from signals? what's their numerical type? what's their dimensionality?) must follow the Mat
format for using the OpenCV implementation of SVM. This is the only requirement.