A method for shape based object classification is presented.Unlike object dimension based methods it does not require any system calibration techniques. A number of 3D object models are utilized as a source of training dataset for a specified camera orientation. Usage of the 3D models allows to perform the dataset creation process semiautomatically. The background subtraction method is used for the purpose of detecting moving objects and Kalman filters based method is utilized for object tracking. Detected objects are parameterized and then classified using a set of SVM classifiers. Probability of each classification attempt is calculated and averaged over object lifetime resulting in effectiveness improvement. The method classification efficiency is tested during experiments for two variouscamera angles and for two various feature vector lengths.
Autorzy
Informacje dodatkowe
- Kategoria
- Aktywność konferencyjna
- Typ
- publikacja w wydawnictwie zbiorowym recenzowanym (także w materiałach konferencyjnych)
- Język
- angielski
- Rok wydania
- 2011
