Title :
A novel system for automatic detection and classification of animal
Author :
Matuska, Slavomir ; Hudec, Robert ; Benco, Miroslav ; Kamencay, Patrik ; Zachariasova, Martina
Author_Institution :
Dept. of Telecommun. & Multimedia, Univ. of Zilina, Zilina, Slovakia
Abstract :
In this paper, a novel system for automatic detection and classification of animal is presented. System called ASFAR (Automatic System For Animal Recognition) is based on distributed so-called `watching device´ in designated area and main computing unit (MCU) acting as server and system manager. Watching devices are situated in wild nature and their task is to detect animal and then send data to MCU to evaluation. The main task of whole system is to determine migration corridors of wild animals in designated area. To create object representation, visual descriptors were chosen and Support Vector Machine (SVM) was used to classify descriptors.
Keywords :
image classification; image representation; object detection; object recognition; support vector machines; ASFAR; automatic animal classification; automatic animal detection; automatic system for animal recognition; distributed watching device; main computing unit; object representation; support vector machine; visual descriptors; wild animal migration corridor determination; Detectors; Feature extraction; Heating; Image segmentation; Roads; Robustness; Training; ASFAR; Support Vector Machine; automatic system; object recognition; visual descriptor;
Conference_Titel :
ELEKTRO, 2014
Conference_Location :
Rajecke Teplice
Print_ISBN :
978-1-4799-3720-2
DOI :
10.1109/ELEKTRO.2014.6847875
