|
|
Re-Identification of Vehicles in Video
Zapletal, Dominik ; Sochor, Jakub (referee) ; Herout, Adam (advisor)
This thesis deals with the vehicle re-identification in video problem. Vehicle re-identification is based on matching image parts obtained from different cameras. This work is focues on the re-identification itself assuming that the vehicle detection problem is already solved including extraction of a full-fledged 3D bounding box. The re-identification problem is solved by using color histograms, histograms of oriented gradients by a linear regressor. The features are used in separate models in order to get the best results in the shortest CPU computation time. The proposed method works with a high accuracy (60% true positives retrieved with 10% false positive rate on a challenging subset of the test data) in 85 milliseconds of the CPU (Core i7) computation time per one vehicle re-identification assuming the Full HD resolution video input. The applications of this work include finding important parameters like travel time, traffic flow, or traffic information in a distributed traffic surveillance and monitoring system.
|
|
|
Image objects detection based on the feature points matching
Trávníček, Vojtěch ; Macíček, Ondřej (referee) ; Harabiš, Vratislav (advisor)
This paper is concerned in branch of computer vision. Methods for extracting feature points are presented as tools for image comparison and finding objects in images. Four methods are metioned which are compared with respect to their effectiveness and utilization. Algorythms SIFT and SURF are described as a state-of-the-arts. This paper also mentions methods for describing feature points and their comparison. Testing images are inserted as a tool for first testing of implemented algorythm. Finally, the implemented method SURF is described and tested with respect to several most relevant parameters.
|
|
|
Image objects detection based on the feature points matching
Trávníček, Vojtěch ; Macíček, Ondřej (referee) ; Harabiš, Vratislav (advisor)
This paper is concerned in branch of computer vision. Methods for extracting feature points are presented as tools for image comparison and finding objects in images. Four methods are metioned which are compared with respect to their effectiveness and utilization. Algorythms SIFT and SURF are described as a state-of-the-arts. This paper also mentions methods for describing feature points and their comparison. Testing images are inserted as a tool for first testing of implemented algorythm. Finally, the implemented method SURF is described and tested with respect to several most relevant parameters.
|
|
|
Re-Identification of Vehicles in Video
Zapletal, Dominik ; Sochor, Jakub (referee) ; Herout, Adam (advisor)
This thesis deals with the vehicle re-identification in video problem. Vehicle re-identification is based on matching image parts obtained from different cameras. This work is focues on the re-identification itself assuming that the vehicle detection problem is already solved including extraction of a full-fledged 3D bounding box. The re-identification problem is solved by using color histograms, histograms of oriented gradients by a linear regressor. The features are used in separate models in order to get the best results in the shortest CPU computation time. The proposed method works with a high accuracy (60% true positives retrieved with 10% false positive rate on a challenging subset of the test data) in 85 milliseconds of the CPU (Core i7) computation time per one vehicle re-identification assuming the Full HD resolution video input. The applications of this work include finding important parameters like travel time, traffic flow, or traffic information in a distributed traffic surveillance and monitoring system.
|
guest ::
