Small Sample Learning of Superpixel Classifiers for EM Segmentation

Pixel and superpixel classifiers have become essential tools for EM segmentation algorithms. Training these classifiers remains a major bottleneck primarily due to the requirement of completely annotating the dataset which is tedious, error-prone and costly. In this paper, we propose an interactive learning scheme for the superpixel classifier for EM segmentation. Our algorithm is 'active semi-supervised' because it requests the labels of a small number of examples from user and applies label propagation technique to generate these queries. Using only a small set (< 20%) of all datapoints, the proposed algorithm consistently generates a classifier almost as accurate as that estimated from a complete groundtruth. We provide segmentation results on multiple datasets to show the strength of these classifiers.

[1]  Anirban Chakraborty,et al.  A Context-aware Delayed Agglomeration Framework for EM Segmentation , 2014, ArXiv.

[2]  Matthew Cook,et al.  Efficient automatic 3D-reconstruction of branching neurons from EM data , 2012, 2012 IEEE Conference on Computer Vision and Pattern Recognition.

[3]  Alexander Zien,et al.  Semi-Supervised Learning , 2006 .

[4]  Charless C. Fowlkes,et al.  Contour Detection and Hierarchical Image Segmentation , 2011, IEEE Transactions on Pattern Analysis and Machine Intelligence.

[5]  H. Sebastian Seung,et al.  Boundary Learning by Optimization with Topological Constraints , 2010, 2010 IEEE Computer Society Conference on Computer Vision and Pattern Recognition.

[6]  Jianbo Shi,et al.  Machine Learning of Hierarchical Clustering to Segment 2D and 3D Images , 2013, PloS one.

[7]  Louis K. Scheffer,et al.  Semi-automated reconstruction of neural circuits using electron microscopy , 2010, Current Opinion in Neurobiology.

[8]  Ronen Basri,et al.  Co-clustering of image segments using convex optimization applied to EM neuronal reconstruction , 2010, 2010 IEEE Computer Society Conference on Computer Vision and Pattern Recognition.

[9]  Gary L. Miller,et al.  A Nearly-m log n Time Solver for SDD Linear Systems , 2011, 2011 IEEE 52nd Annual Symposium on Foundations of Computer Science.

[10]  H. Sebastian Seung,et al.  Learning to Agglomerate Superpixel Hierarchies , 2011, NIPS.

[11]  Louis K. Scheffer,et al.  A visual motion detection circuit suggested by Drosophila connectomics , 2013, Nature.

[12]  Leo Breiman,et al.  Random Forests , 2001, Machine Learning.

[13]  J. Lafferty,et al.  Combining active learning and semi-supervised learning using Gaussian fields and harmonic functions , 2003, ICML 2003.

[14]  John Langford,et al.  Importance weighted active learning , 2008, ICML '09.

[15]  Craig A. Knoblock,et al.  Active + Semi-supervised Learning = Robust Multi-View Learning , 2002, ICML.

[16]  Ullrich Köthe,et al.  Segmentation of SBFSEM Volume Data of Neural Tissue by Hierarchical Classification , 2008, DAGM-Symposium.

[17]  Ullrich Köthe,et al.  Ilastik: Interactive learning and segmentation toolkit , 2011, 2011 IEEE International Symposium on Biomedical Imaging: From Nano to Macro.

[18]  Joachim M. Buhmann,et al.  Neuron geometry extraction by perceptual grouping in ssTEM images , 2010, 2010 IEEE Computer Society Conference on Computer Vision and Pattern Recognition.

[19]  Ullrich Köthe,et al.  Globally Optimal Closed-Surface Segmentation for Connectomics , 2012, ECCV.

[20]  Shang-Hua Teng,et al.  A Local Clustering Algorithm for Massive Graphs and Its Application to Nearly Linear Time Graph Partitioning , 2008, SIAM J. Comput..

[21]  M. Helmstaedter.  Cellular-resolution connectomics: challenges of dense neural circuit reconstruction , 2013, Nature Methods.

[22]  Xiaojin Zhu,et al.  Introduction to Semi-Supervised Learning , 2009, Synthesis Lectures on Artificial Intelligence and Machine Learning.