Extrakce významných rysů v obrazových scénách s využitím v klasifikačních technikách

Abstract

In this thesis, we explore image feature extraction approach called Scale-Invariant Feature Transform. We use the extracted features for training and testing of an image classification technique, a Support Vector Machine. In first few sections, we introduce the Scale-invariant Feature Transform, techniques used to adapt the the feature descriptors to the classifier, and the Support Vector Machine. For the benchmarks, we train and test the classifier using features extracted from a real image data. For comparison, we train and test the classifier using pixel values of the data.