GLAN: A Graph-based Linear Assignment Network

Differentiable solvers for the linear assignment problem (LAP) have attracted much research attention in recent years, which are usually embedded into learning frameworks as components. However, previous algorithms, with or without learning strategies, usually suffer from the degradation of the optimality with the increment of the problem size. In this paper, we propose a learnable linear assignment solver based on deep graph networks. Specifically, we first transform the cost matrix to a bipartite graph and convert the assignment task to the problem of selecting reliable edges from the constructed graph. Subsequently, a deep graph network is developed to aggregate and update the features of nodes and edges. Finally, the network predicts a label for each edge that indicates the assignment relationship. The experimental results on a synthetic dataset reveal that our method outperforms state-of-the-art baselines and achieves consistently high accuracy with the increment of the problem size. Furthermore, we also embed the proposed solver, in comparison with state-of-the-art baseline solvers, into a popular multi-object tracking (MOT) framework to train the tracker in an end-to-end manner. The experimental results on MOT benchmarks illustrate that the proposed LAP solver improves the tracker by the largest margin.

[1]  Dacheng Tao,et al.  Graph Edge Convolutional Neural Networks for Skeleton-Based Action Recognition , 2018, IEEE Transactions on Neural Networks and Learning Systems.

[2]  Michael I. Jordan Serial Order: A Parallel Distributed Processing Approach , 1997 .

[3]  Stefan Roth,et al.  MOT16: A Benchmark for Multi-Object Tracking , 2016, ArXiv.

[4]  A. Volgenant,et al.  A shortest augmenting path algorithm for dense and sparse linear assignment problems , 1987, Computing.

[5]  Kaiming He,et al.  Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks , 2015, IEEE Transactions on Pattern Analysis and Machine Intelligence.

[6]  Sanja Fidler,et al.  DMM-Net: Differentiable Mask-Matching Network for Video Object Segmentation , 2019, 2019 IEEE/CVF International Conference on Computer Vision (ICCV).

[7]  D. Bertsekas The auction algorithm: A distributed relaxation method for the assignment problem , 1988 .

[8]  M. Trick A Linear Relaxation Heuristic for the Generalized Assignment Problem , 1992 .

[9]  R. Dykstra,et al.  A Method for Finding Projections onto the Intersection of Convex Sets in Hilbert Spaces , 1986 .

[10]  Philip S. Yu,et al.  A Comprehensive Survey on Graph Neural Networks , 2019, IEEE Transactions on Neural Networks and Learning Systems.

[11]  Marco Cuturi,et al.  Sinkhorn Distances: Lightspeed Computation of Optimal Transport , 2013, NIPS.

[12]  Yong-Hwa Kim,et al.  Bidirectional Long Short-Term Memory Neural Networks for Linear Sum Assignment Problems , 2019 .

[13]  Mohammed El-Beltagy,et al.  On the Optimality and Speed of the Deep Greedy Switching Algorithm for Linear Assignment Problems , 2013, 2013 IEEE International Symposium on Parallel & Distributed Processing, Workshops and Phd Forum.

[14]  Xavier Alameda-Pineda,et al.  How to Train Your Deep Multi-Object Tracker , 2020, 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR).

[15]  Frank D. Julca-Aguilar,et al.  Matching based ground-truth annotation for online handwritten mathematical expressions , 2015, Pattern Recognit..

[16]  Yangyang Xu,et al.  Transductive Zero-Shot Action Recognition via Visually Connected Graph Convolutional Networks , 2020, IEEE Transactions on Neural Networks and Learning Systems.

[17]  Rainer E. Burkard,et al.  Linear Assignment Problems and Extensions , 1999, Handbook of Combinatorial Optimization.

[18]  Narendra Karmarkar,et al.  A new polynomial-time algorithm for linear programming , 1984, Comb..

[19]  Qingshan Liu,et al.  A one-layer projection neural network for linear assignment problem , 2015, 2015 34th Chinese Control Conference (CCC).

[20]  R. Dykstra An Algorithm for Restricted Least Squares Regression , 1983 .

[21]  Laura Leal-Taixé,et al.  Tracking Without Bells and Whistles , 2019, 2019 IEEE/CVF International Conference on Computer Vision (ICCV).

[22]  Yidong Li,et al.  Learning Combinatorial Solver for Graph Matching , 2020, 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR).

[23]  J. Munkres ALGORITHMS FOR THE ASSIGNMENT AND TRANSIORTATION tROBLEMS* , 1957 .

[24]  Jason Altschuler,et al.  Near-linear time approximation algorithms for optimal transport via Sinkhorn iteration , 2017, NIPS.

[25]  Geoffrey Ye Li,et al.  Deep Neural Networks for Linear Sum Assignment Problems , 2018, IEEE Wireless Communications Letters.

[26]  장윤희,et al.  Y. , 2003, Industrial and Labor Relations Terms.

[27]  L. Livi,et al.  Hierarchical Representation Learning in Graph Neural Networks With Node Decimation Pooling , 2019, IEEE Transactions on Neural Networks and Learning Systems.

[28]  Andrew V. Goldberg,et al.  An efficient cost scaling algorithm for the assignment problem , 1995, Math. Program..

[29]  Mauricio G. C. Resende,et al.  Greedy Randomized Adaptive Search Procedures , 1995, J. Glob. Optim..

[30]  M. Akgül,et al.  A dual feasible forest algorithm for the linear assignment problem , 1991 .

[31]  Lei Zhang,et al.  A Concave Optimization Algorithm for Matching Partially Overlapping Point Sets , 2017, Pattern Recognit..

[32]  Mubarak Shah,et al.  Deep Affinity Network for Multiple Object Tracking , 2018, IEEE Transactions on Pattern Analysis and Machine Intelligence.

[33]  Yoshua Bengio,et al.  Convolutional networks for images, speech, and time series , 1998 .

[34]  K. G. Ramakrishnan,et al.  An Approximate Dual Projective Algorithm for Solving Assignment Problems , 1991, Network Flows And Matching.

[35]  K. G. Ramakrishnan,et al.  Computational results of an interior point algorithm for large scale linear programming , 1991, Math. Program..

[36]  Richard Sinkhorn A Relationship Between Arbitrary Positive Matrices and Doubly Stochastic Matrices , 1964 .

[37]  Harold W. Kuhn,et al.  The Hungarian method for the assignment problem , 1955, 50 Years of Integer Programming.