TY - JOUR
T1 - A Deep Learning Framework for Joint Image Restoration and Recognition
AU - Chen, Ruilong
AU - Mihaylova, Lyudmila
AU - Zhu, Hao
AU - Bouaynaya, Nidhal Carla
N1 - Publisher Copyright:
© 2019, The Author(s).
PY - 2020/3/1
Y1 - 2020/3/1
N2 - Image restoration and recognition are important computer vision tasks representing an inherent part of autonomous systems. These two tasks are often implemented in a sequential manner, in which the restoration process is followed by a recognition. In contrast, this paper proposes a joint framework that simultaneously performs both tasks within a shared deep neural network architecture. This joint framework integrates the restoration and recognition tasks by incorporating: (i) common layers, (ii) restoration layers and (iii) classification layers. The total loss function combines the restoration and classification losses. The proposed joint framework, based on capsules, provides an efficient solution that can cope with challenges due to noise, image rotations and occlusions. The developed framework has been validated and evaluated on a public vehicle logo dataset under various degradation conditions, including Gaussian noise, rotation and occlusion. The results show that the joint framework improves the accuracy compared with the single task networks.
AB - Image restoration and recognition are important computer vision tasks representing an inherent part of autonomous systems. These two tasks are often implemented in a sequential manner, in which the restoration process is followed by a recognition. In contrast, this paper proposes a joint framework that simultaneously performs both tasks within a shared deep neural network architecture. This joint framework integrates the restoration and recognition tasks by incorporating: (i) common layers, (ii) restoration layers and (iii) classification layers. The total loss function combines the restoration and classification losses. The proposed joint framework, based on capsules, provides an efficient solution that can cope with challenges due to noise, image rotations and occlusions. The developed framework has been validated and evaluated on a public vehicle logo dataset under various degradation conditions, including Gaussian noise, rotation and occlusion. The results show that the joint framework improves the accuracy compared with the single task networks.
UR - http://www.scopus.com/inward/record.url?scp=85070239281&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85070239281&partnerID=8YFLogxK
U2 - 10.1007/s00034-019-01222-x
DO - 10.1007/s00034-019-01222-x
M3 - Article
AN - SCOPUS:85070239281
SN - 0278-081X
VL - 39
SP - 1561
EP - 1580
JO - Circuits, Systems, and Signal Processing
JF - Circuits, Systems, and Signal Processing
IS - 3
ER -