A novel contextual memory algorithm for edge detection

Edge detection plays an important role in many computer vision systems. In this paper, we propose a novel application agnostic algorithm for prediction of probabilities based on the contextual information available and then apply the algorithm for estimating the probability of pixels belonging to an edge using surrounding pixel values as local contexts. We then proceed to test different image transformations as input layers, such as the Canny edge detector. We propose two different architectures, one single layered and one multilayered, which approach the scaling problem by creating scaled side outputs and combining them via a logistic regression layer. We tested our approach on the BSDS500 edge detection dataset with optimistic results.

[1]  Xin Yao,et al.  Ensemble learning via negative correlation , 1999, Neural Networks.

[2]  Christopher Mattern.  Mixing Strategies in Data Compression , 2012, 2012 Data Compression Conference.

[3]  Jinhui Tang,et al.  Richer Convolutional Features for Edge Detection , 2017, IEEE Transactions on Pattern Analysis and Machine Intelligence.

[4]  T. Minka.  A comparison of numerical optimizers for logistic regression , 2004 .

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

[6]  Seyed-Ahmad Ahmadi,et al.  V-Net: Fully Convolutional Neural Networks for Volumetric Medical Image Segmentation , 2016, 2016 Fourth International Conference on 3D Vision (3DV).

[7]  Nathan Intrator,et al.  Optimal ensemble averaging of neural networks , 1997 .

[8]  Yu Liu,et al.  Learning Relaxed Deep Supervision for Better Edge Detection , 2016, 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR).

[9]  C. Lawrence Zitnick,et al.  Fast Edge Detection Using Structured Forests , 2014, IEEE Transactions on Pattern Analysis and Machine Intelligence.

[10]  Nicu Sebe,et al.  Learning Deep Structured Multi-Scale Features using Attention-Gated CRFs for Contour Prediction , 2017, NIPS.

[11]  Xiang Bai,et al.  Richer Convolutional Features for Edge Detection , 2016, IEEE Transactions on Pattern Analysis and Machine Intelligence.

[12]  Rocco A. Servedio,et al.  Random classification noise defeats all convex potential boosters , 2008, ICML '08.

[13]  Ming-Yu Liu,et al.  CASENet: Deep Category-Aware Semantic Edge Detection , 2017, 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR).

[14]  Honglak Lee,et al.  Object Contour Detection with a Fully Convolutional Encoder-Decoder Network , 2016, 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR).

[15]  Bilwaj Gaonkar,et al.  Deep learning in the small sample size setting: cascaded feed forward neural networks for medical image segmentation , 2016, SPIE Medical Imaging.

[16]  Matthew V. Mahoney,et al.  Adaptive weighing of context models for lossless data compression , 2005 .

[17]  Hua Fan,et al.  An improved adaptive edge detection algorithm based on Canny , 2018, Other Conferences.

[18]  Edward S. Deutsch,et al.  On the Quantitative Evaluation of Edge Detection Schemes and their Comparison with Human Performance , 1975, IEEE Transactions on Computers.

[19]  Kaiqi Huang,et al.  Deep Crisp Boundaries: From Boundaries to Higher-Level Tasks , 2018, IEEE Transactions on Image Processing.

[20]  Javier Montero,et al.  A novel edge detection algorithm based on a hierarchical graph-partition approach , 2018, J. Intell. Fuzzy Syst..