TY - GEN
T1 - Humanoid robot detection using deep learning
T2 - 21st RoboCup International Symposium, 2017
AU - Javadi, Mohammad
AU - Azar, Sina Mokhtarzadeh
AU - Azami, Sajjad
AU - Ghidary, Saeed Shiry
AU - Sadeghnejad, Soroush
AU - Baltes, Jacky
N1 - Publisher Copyright:
© Springer Nature Switzerland AG 2018.
PY - 2018
Y1 - 2018
N2 - Recent advances in computer vision have made the detection of landmarks on the soccer field easier for teams. However, the detection of other robots is also a critical capability that has not garnered much attention in the RoboCup community so far. This problem is well represented in different RoboCup Soccer and Rescue Robot Leagues. In this paper, we compare several two-stage detection systems based on various Convolutional Neural Networks (CNN) and highlight their speed-accuracy trade off. The approach performs edge based image segmentation in order to reduce the search space and then a CNN validates the detection in the second stage. We use images of different humanoid robots to train and test three different CNN architectures. A part of these images was gathered by our team and will be publicly available. Our experiments demonstrate the strong adaptability of deeper CNNs. These models, trained on a limited set of robots, are able to successfully distinguish an unseen kind of humanoid robot from non-robot regions.
AB - Recent advances in computer vision have made the detection of landmarks on the soccer field easier for teams. However, the detection of other robots is also a critical capability that has not garnered much attention in the RoboCup community so far. This problem is well represented in different RoboCup Soccer and Rescue Robot Leagues. In this paper, we compare several two-stage detection systems based on various Convolutional Neural Networks (CNN) and highlight their speed-accuracy trade off. The approach performs edge based image segmentation in order to reduce the search space and then a CNN validates the detection in the second stage. We use images of different humanoid robots to train and test three different CNN architectures. A part of these images was gathered by our team and will be publicly available. Our experiments demonstrate the strong adaptability of deeper CNNs. These models, trained on a limited set of robots, are able to successfully distinguish an unseen kind of humanoid robot from non-robot regions.
KW - Convolutional neural networks
KW - Deep learning
KW - Humanoid robots
KW - Image segmentation
KW - Robot detection
KW - Robot vision
UR - http://www.scopus.com/inward/record.url?scp=85053941550&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85053941550&partnerID=8YFLogxK
U2 - 10.1007/978-3-030-00308-1_28
DO - 10.1007/978-3-030-00308-1_28
M3 - Conference contribution
AN - SCOPUS:85053941550
SN - 9783030003074
T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
SP - 338
EP - 349
BT - RoboCup 2017
A2 - Akiyama, Hidehisa
A2 - Obst, Oliver
A2 - Sammut, Claude
A2 - Tonidandel, Flavio
PB - Springer Verlag
Y2 - 27 July 2017 through 31 July 2017
ER -