Pregled bibliografske jedinice broj: 1031286
Detection of Toy Soldiers Taken from a Bird’s Perspective Using Convolutional Neural Networks
Detection of Toy Soldiers Taken from a Bird’s Perspective Using Convolutional Neural Networks // 11th International Conference ICT Innovations / Gievska, S. ; Madjarov, G. (ur.).
Ohrid, Sjeverna Makedonija: Springer, 2019. str. 13-26 doi:10.1007/978-3-030-33110-8_2 (poster, međunarodna recenzija, cjeloviti rad (in extenso), znanstveni)
CROSBI ID: 1031286 Za ispravke kontaktirajte CROSBI podršku putem web obrasca
Naslov
Detection of Toy Soldiers Taken from a Bird’s
Perspective Using Convolutional Neural Networks
Autori
Sambolek, Saša ; Ivašić-Kos, Marina
Vrsta, podvrsta i kategorija rada
Radovi u zbornicima skupova, cjeloviti rad (in extenso), znanstveni
ISBN
978-3-030-33109-2
Skup
11th International Conference ICT Innovations
Mjesto i datum
Ohrid, Sjeverna Makedonija, 17.10.2019. - 19.10.2019
Vrsta sudjelovanja
Poster
Vrsta recenzije
Međunarodna recenzija
Ključne riječi
Object detectors ; SSD ; Faster R-CNN
Sažetak
This paper describes the use of two different deep-learning approaches for object detection to recognize a toy soldier. We use recordings of toy soldiers in different poses under different scenarios to simulate the appearance of persons on footage taken by drones. Recordings from a bird’s eye view are today widely used in the search for missing persons in non-urban areas, border control, animal movement control, and the like. We have compared the single-shot multi-box detector (SSD) with the MobileNet or Inception V2 as a backbone, SSDLite with MobileNet and Faster R- CNN combined with Inception V2 and ResNet50. The results show that Faster R-CNN detects small objects such as toy soldiers more successfully than SSD, and the training time of Faster R-CNN is much shorter than that of SSD.
Izvorni jezik
Engleski
Znanstvena područja
Računarstvo, Informacijske i komunikacijske znanosti
POVEZANOST RADA
Projekti:
HRZZ-IP-2016-06-8345 - Automatsko raspoznavanje akcija i aktivnosti u multimedijalnom sadržaju iz domene sporta (RAASS) (Ivašić Kos, Marina, HRZZ - 2016-06) ( CroRIS)
Ustanove:
Fakultet informatike i digitalnih tehnologija, Rijeka
Citiraj ovu publikaciju:
Časopis indeksira:
- Web of Science Core Collection (WoSCC)
- Conference Proceedings Citation Index - Science (CPCI-S)
- Scopus