Pregled bibliografske jedinice broj: 890888
I Know That Person: Generative Full Body and Face De-identification of People in Images
I Know That Person: Generative Full Body and Face De-identification of People in Images // Computer Vision and Pattern Recognition Workshops (CVPRW), 2017 IEEE Conference on
Honolulu (HI), Sjedinjene Američke Države: Institute of Electrical and Electronics Engineers (IEEE), 2017. str. 1319-1328 (poster, međunarodna recenzija, cjeloviti rad (in extenso), znanstveni)
CROSBI ID: 890888 Za ispravke kontaktirajte CROSBI podršku putem web obrasca
Naslov
I Know That Person: Generative Full Body and Face De-identification of People in Images
Autori
Brkić, Karla ; Sikirić, Ivan ; Hrkać, Tomislav ; Kalafatić, Zoran
Vrsta, podvrsta i kategorija rada
Radovi u zbornicima skupova, cjeloviti rad (in extenso), znanstveni
Izvornik
Computer Vision and Pattern Recognition Workshops (CVPRW), 2017 IEEE Conference on
/ - : Institute of Electrical and Electronics Engineers (IEEE), 2017, 1319-1328
Skup
The First International Workshop on The Bright and Dark Sides of Computer Vision: Challenges and Opportunities for Privacy and Security (CV- COPS 2017)
Mjesto i datum
Honolulu (HI), Sjedinjene Američke Države, 21.07.2017
Vrsta sudjelovanja
Poster
Vrsta recenzije
Međunarodna recenzija
Ključne riječi
computer vision, de-identification
Sažetak
We propose a model for full body and face de- identification of humans in images. Given a segmentation of the human figure, our model generates a synthetic human image with an alternative appearance that looks natural and fits the segmentation outline. The model is usable with various levels of segmentation, from simple human figure blobs to complex garment-level segmentations. The level of detail in the de-identified output depends on the level of detail in the input segmentation. The model de-identifies not only primary biometric identifiers (e.g. the face), but also soft and non-biometric identifiers including clothing, hairstyle, etc. Quantitative and perceptual experiments indicate that our model produces de-identified outputs that thwart human and machine recognition, while preserving data utility and naturalness.
Izvorni jezik
Engleski
POVEZANOST RADA
Ustanove:
Fakultet elektrotehnike i računarstva, Zagreb