Deep Learning Improves Template Matching by Normalized Cross Correlation

Template matching by normalized cross correlation (NCC) is widely used for finding image correspondences. We improve the robustness of this algorithm by preprocessing images with "siamese" convolutional networks trained to maximize the contrast between NCC values of true and false matches. The improvement is quantified using patches of brain images from serial section electron microscopy. Relative to a parameter-tuned bandpass filter, siamese convolutional networks significantly reduce false matches. Furthermore, all false matches can be eliminated by removing a tiny fraction of all matches based on NCC values. The improved accuracy of our method could be essential for connectomics, because emerging petascale datasets may require billions of template matches to assemble 2D images of serial sections into a 3D image stack. Our method is also expected to generalize to many other computer vision applications that use NCC template matching to find image correspondences.

[1]  Yann LeCun,et al.  Signature Verification Using A "Siamese" Time Delay Neural Network , 1993, Int. J. Pattern Recognit. Artif. Intell..

[2]  Jitendra Malik,et al.  Geometric blur for template matching , 2001, Proceedings of the 2001 IEEE Computer Society Conference on Computer Vision and Pattern Recognition. CVPR 2001.

[3]  Yann LeCun,et al.  Learning a similarity metric discriminatively, with application to face verification , 2005, 2005 IEEE Computer Society Conference on Computer Vision and Pattern Recognition (CVPR'05).

[4]  Rong Jin,et al.  Distance Metric Learning: A Comprehensive Survey , 2006 .

[5]  Brian B. Avants,et al.  Symmetric diffeomorphic image registration with cross-correlation: Evaluating automated labeling of elderly and neurodegenerative brain , 2008, Medical Image Anal..

[6]  J. P. Lewis,et al.  Fast Template Matching , 2009 .

[7]  Torsten Rohlfing,et al.  Bead-based mosaicing of single plane illumination microscopy images using geometric local descriptor matching , 2009, Medical Imaging.

[8]  Jianwen Luo,et al.  A fast normalized cross-correlation calculation method for motion estimation , 2010, IEEE Transactions on Ultrasonics, Ferroelectrics and Frequency Control.

[9]  Sang Uk Lee,et al.  Robust Stereo Matching Using Adaptive Normalized Cross-Correlation , 2011, IEEE Transactions on Pattern Analysis and Machine Intelligence.

[10]  A. Cardona,et al.  Elastic volume reconstruction from series of ultra-thin microscopy sections , 2012, Nature Methods.

[11]  Brian Kulis,et al.  Metric Learning: A Survey , 2013, Found. Trends Mach. Learn..

[12]  Trevor Darrell,et al.  Do Convnets Learn Correspondence? , 2014, NIPS.

[13]  Simone Calderara,et al.  Visual Tracking: An Experimental Survey , 2014, IEEE Transactions on Pattern Analysis and Machine Intelligence.

[14]  Nir Shavit,et al.  The big data challenges of connectomics , 2014, Nature Neuroscience.

[15]  Thomas Brox,et al.  FlowNet: Learning Optical Flow with Convolutional Networks , 2015, 2015 IEEE International Conference on Computer Vision (ICCV).

[16]  Anurag Mittal,et al.  Deep Neural Networks with Inexact Matching for Person Re-Identification , 2016, NIPS.

[17]  Reza Bosagh Zadeh,et al.  FusionNet: 3D Object Classification Using Multiple Data Representations , 2016, ArXiv.

[18]  Trevor Darrell,et al.  Learning Features by Watching Objects Move , 2016, 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR).