From concrete to abstract: Multilayer neural networks for disaster victims detection

Search-and-rescue (SAR) team main objective is to quickly locate victims in post-disaster scenario. In such disaster scenario, images are usually complex containing highly cluttered background such as debris, soil, gravel, ruined building, and clothes, which are difficult to distinguish from the victims. Previous methods which only work on nearly uniform background taken from either indoor or yard are not suitable and can deteriorate the detection system. In this paper, we demonstrate the feasibility of multilayer neural network for disaster victims detection on highly cluttered background. Theoretical justification from which deep learning learns from concrete to object abstraction is established. In order to build a more discriminative system, this theoretical justification then leads us to perform pretraining using data-rich datasheet followed by finetuning only on the last layers using data-specific datasheet while keeping the other layers fixed. A new Indonesian disaster victims datasheet is also provided. Experimental results show the efficiency of the method for disaster victims detection in highly cluttered background.

[1]  Trevor Darrell,et al.  Rich Feature Hierarchies for Accurate Object Detection and Semantic Segmentation , 2013, 2014 IEEE Conference on Computer Vision and Pattern Recognition.

[2]  W.E. Green,et al.  A competition to identify key challenges for unmanned aerial robots in near-earth environments , 2005, ICAR '05. Proceedings., 12th International Conference on Advanced Robotics, 2005..

[3]  Rob Fergus,et al.  Visualizing and Understanding Convolutional Networks , 2013, ECCV.

[4]  Maja J. Mataric,et al.  A laser-based people tracker , 2002, Proceedings 2002 IEEE International Conference on Robotics and Automation (Cat. No.02CH37292).

[5]  Luc Van Gool,et al.  The Pascal Visual Object Classes (VOC) Challenge , 2010, International Journal of Computer Vision.

[6]  Alexander Carballo,et al.  Multiple People Detection from a Mobile Robot using Double Layered Laser Range Finders , 2009 .

[7]  Andrew Zisserman,et al.  Very Deep Convolutional Networks for Large-Scale Image Recognition , 2014, ICLR.

[8]  Fawzi Nashashibi,et al.  Centralized fusion for fast people detection in dense environment , 2009, 2009 IEEE International Conference on Robotics and Automation.

[9]  Patrick Sayd,et al.  Real-Time Posture Analysis in a Crowd using Thermal Imaging , 2007, 2007 IEEE Conference on Computer Vision and Pattern Recognition.

[10]  Michael A. Goodrich,et al.  Towards combining UAV and sensor operator roles in UAV-enabled visual search , 2008, 2008 3rd ACM/IEEE International Conference on Human-Robot Interaction (HRI).

[11]  Patrick Doherty,et al.  A UAV Search and Rescue Scenario with Human Body Detection and Geolocalization , 2007, Australian Conference on Artificial Intelligence.

[12]  Rogelio Lozano,et al.  Human detection in uncluttered environments: From ground to UAV view , 2014, 2014 13th International Conference on Control Automation Robotics & Vision (ICARCV).

[13]  Bernt Schiele,et al.  Vision based victim detection from unmanned aerial vehicles , 2010, 2010 IEEE/RSJ International Conference on Intelligent Robots and Systems.

[14]  Koen E. A. van de Sande,et al.  Selective Search for Object Recognition , 2013, International Journal of Computer Vision.

[15]  Patrick Doherty,et al.  Vision for a UAV helicopter , 2002 .

[16]  James W. Davis,et al.  Robust detection of people in thermal imagery , 2004, Proceedings of the 17th International Conference on Pattern Recognition, 2004. ICPR 2004..

[17]  Wolfram Burgard,et al.  Efficient people tracking in laser range data using a multi-hypothesis leg-tracker with adaptive occlusion probabilities , 2008, 2008 IEEE International Conference on Robotics and Automation.

[18]  Jemal H. Abawajy,et al.  Text Detection in Low Resolution Scene Images Using Convolutional Neural Network , 2016, SCDM.

[19]  Rogelio Lozano,et al.  Fast and viewpoint robust human detection for SAR operations , 2014, 2014 IEEE International Symposium on Safety, Security, and Rescue Robotics (2014).

[20]  Ross B. Girshick,et al.  Fast R-CNN , 2015, 1504.08083.

[21]  A. Broggi,et al.  Pedestrian Detection using Infrared images and Histograms of Oriented Gradients , 2006, 2006 IEEE Intelligent Vehicles Symposium.

[22]  Geoffrey E. Hinton,et al.  ImageNet classification with deep convolutional neural networks , 2012, Commun. ACM.

[23]  Trevor Darrell,et al.  Caffe: Convolutional Architecture for Fast Feature Embedding , 2014, ACM Multimedia.

[24]  Wolfram Burgard,et al.  People Tracking with Mobile Robots Using Sample-Based Joint Probabilistic Data Association Filters , 2003, Int. J. Robotics Res..

[25]  Jitendra Malik,et al.  Analyzing the Performance of Multilayer Neural Networks for Object Recognition , 2014, ECCV.

[26]  Sang Joon Kim,et al.  A Mathematical Theory of Communication , 2006 .

[27]  Wolfram Burgard,et al.  Using Boosted Features for the Detection of People in 2D Range Data , 2007, Proceedings 2007 IEEE International Conference on Robotics and Automation.

[28]  Alexander Kleiner,et al.  Genetic MRF model optimization for real-time victim detection in search and rescue , 2007, 2007 IEEE/RSJ International Conference on Intelligent Robots and Systems.