TY - GEN
T1 - Curriculum learning for multi-task classification of visual attributes
AU - Sarafianos, Nikolaos
AU - Giannakopoulos, Theodore
AU - Nikou, Christophoros
AU - Kakadiaris, Ioannis A.
N1 - Publisher Copyright:
© 2017 IEEE.
PY - 2017/7/1
Y1 - 2017/7/1
N2 - Visual attributes, from simple objects (e.g., backpacks, hats) to soft-biometrics (e.g., gender, height, clothing) have proven to be a powerful representational approach for many applications such as image description and human identification. In this paper, we introduce a novel method to combine the advantages of both multi-task and curriculum learning in a visual attribute classification framework. Individual tasks are grouped based on their correlation so that two groups of strongly and weakly correlated tasks are formed. The two groups of tasks are learned in a curriculum learning setup by transferring the acquired knowledge from the strongly to the weakly correlated. The learning process within each group though, is performed in a multitask classification setup. The proposed method learns better and converges faster than learning all the tasks in a typical multi-task learning paradigm. We demonstrate the effectiveness of our approach on the publicly available, SoBiR, VIPeR and PETA datasets and report state-of-the-art results across the board.
AB - Visual attributes, from simple objects (e.g., backpacks, hats) to soft-biometrics (e.g., gender, height, clothing) have proven to be a powerful representational approach for many applications such as image description and human identification. In this paper, we introduce a novel method to combine the advantages of both multi-task and curriculum learning in a visual attribute classification framework. Individual tasks are grouped based on their correlation so that two groups of strongly and weakly correlated tasks are formed. The two groups of tasks are learned in a curriculum learning setup by transferring the acquired knowledge from the strongly to the weakly correlated. The learning process within each group though, is performed in a multitask classification setup. The proposed method learns better and converges faster than learning all the tasks in a typical multi-task learning paradigm. We demonstrate the effectiveness of our approach on the publicly available, SoBiR, VIPeR and PETA datasets and report state-of-the-art results across the board.
UR - http://www.scopus.com/inward/record.url?scp=85046291817&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85046291817&partnerID=8YFLogxK
U2 - 10.1109/ICCVW.2017.306
DO - 10.1109/ICCVW.2017.306
M3 - Conference contribution
AN - SCOPUS:85046291817
T3 - Proceedings - 2017 IEEE International Conference on Computer Vision Workshops, ICCVW 2017
SP - 2608
EP - 2615
BT - Proceedings - 2017 IEEE International Conference on Computer Vision Workshops, ICCVW 2017
PB - Institute of Electrical and Electronics Engineers Inc.
T2 - 16th IEEE International Conference on Computer Vision Workshops, ICCVW 2017
Y2 - 22 October 2017 through 29 October 2017
ER -