Název: | Isomorphic loss function for head pose estimation |
Autoři: | Felea, Iulian Florea, Corneliu Vertan, Constantin Florea, Laura |
Citace zdrojového dokumentu: | WSCG 2017: poster papers proceedings: 25th International Conference in Central Europe on Computer Graphics, Visualization and Computer Visionin co-operation with EUROGRAPHICS Association, p. 89-94. |
Datum vydání: | 2017 |
Nakladatel: | Václav Skala - UNION Agency |
Typ dokumentu: | konferenční příspěvek conferenceObject |
URI: | wscg.zcu.cz/WSCG2017/!!_CSRN-2703.pdf http://hdl.handle.net/11025/29618 |
ISBN: | 978-80-86943-46-6 |
ISSN: | 2464-4617 |
Klíčová slova: | odhad pozice hlavy;funkce ztráty;vzdálenost L1;izomorfní transformace |
Klíčová slova v dalším jazyce: | head pose estimation;loss function;L1 distance;isomorphic transformation |
Abstrakt: | Accurate head pose estimation is a key step in many practical applications involving face analysis tasks, such as emotion recognition. We address the problem of head pose estimation in still color images acquired with a standard camera with limited resolution details. To achieve the proposed goal, we make use of the recent advances of Deep Convolutional Neural Networks. As head angles with respect on yaw and pitch are continuous, the problem is one of regression. Typical loss function for regression are based on L1 and L2 distances which are notorious for susceptibility to outliers. To address this aspect we introduce an isomorphic transformation which maps the initially infinite space into a closed space compressed at the ends and thus significantly down–weighting the significance of outliers. We have thoroughly evaluated the proposed approach on multiple publicly head pose databases. |
Práva: | © Václav Skala - Union Agency |
Vyskytuje se v kolekcích: | WSCG 2017: Poster Papers Proceedings |
Soubory připojené k záznamu:
Soubor | Popis | Velikost | Formát | |
---|---|---|---|---|
Felea.pdf | Plný text | 1,11 MB | Adobe PDF | Zobrazit/otevřít |
Použijte tento identifikátor k citaci nebo jako odkaz na tento záznam:
http://hdl.handle.net/11025/29618
Všechny záznamy v DSpace jsou chráněny autorskými právy, všechna práva vyhrazena.