Structure of thesis

Chapter 2 gives a short overview of the state of the art, we study the existing methods for vehicle type recognition. We observe two major approaches: model based approach and appearance based approach. It is not our intent to give a complete overview of all of these methods, we rather focus on articles which influenced this work.

Chapter 3 describes the proposed method. We can read here how the region of interest is extracted from the image, how the image is geometrically and photometrically normalized, how the feature vector is extracted from the sample image and what classifiers we used.

Chapter 4 describes the experiments we performed. The chapter starts with database description and continues with a description of parameters used in experiments, i.e. SIFT representations, classifiers and error functions used.

Chapter 5 validates the recognition system we developed with all of its parameters. The recognition system is evaluated against image blur, image added noise and learning set reduction.

Chapter 6 presents the conclusions of our work.