A vocabulary tree for image classification have been designed to be integrated in
mobile robotic applications. It is a learning schema based on decission trees, bags of
features and inverted files. The design provides training and optimization parameters that
have been characterized using several detectors and descriptors for several input datasets.
Evaluation tests performed on public image databases allow to compare obtained results
with previously published literature. All the tools and resources used in this project are
Open Source licensed.
Be the first to post a review of PyVocabularyTree!