Khondaker Tasrif Noor1, Antonio Robles-Kelly1,

Transcript:Khondaker Tasrif Noor1, Antonio Robles-Kelly1,:
Khondaker Tasrif Noor1, Antonio Robles-Kelly1, Brano Kusy2 1 Deakin University, School of IT, Waurn Ponds, VIC, Australia. 2 CSIRO Data 61, Queensland Centre for Advanced Technologies, QLD, Australia. A Capsule Network for Hierarchical Multi-Label Image Classification IAPR Joint International Workshops on Statistical Techniques in Pattern Recognition (SPR 2022) and Structural and Syntactic Pattern Recognition (SSPR 2022) Contents 2 Introduction 3 Tree-based Hierarchical Structure Figure: Sample hierarchical label tree for CIFAR-10 dataset. 4 Hierarchical Classifiers 5 CNN Based Classifiers 6 Figure: B-CNN1 architecture. Global Hierarchical Classifier Branch CNN (B-CNN1) Relates Hierarchical levels to Conv layers Separate prediction layer Global classification approach X. Zhu and M. Bain. B-cnn: branch convolutional neural network for hierarchical classification. arXiv preprint arXiv:1709.09890, 2017. Capsule Networks 7 S. Sabour, N. Frosst, and G. E. Hinton. Dynamic routing between capsules. Advances in neural information processing systems, 30, 2017. Capsule Networks 8 S. Sabour, N. Frosst, and G. E. Hinton. Dynamic routing between capsules. Advances in neural information processing systems, 30, 2017. Multi-label Capsule Network (ML-CapsNet) 9 Figure: Architecture of multi-label capsule network (ML-CapsNet). Loss Function 10 Results and Discussion 11 Traditional Models: B-CNN1 : CNN based tree architecture. Global hierarchical classifier. CapsNet2 : Not a hierarchical model Flat Classifier Datasets: X. Zhu and M. Bain. B-cnn: branch convolutional neural network for hierarchical classification. arXiv preprint arXiv:1709.09890, 2017. S. Sabour, N. Frosst, and G. E. Hinton. Dynamic routing between capsules. Advances in neural information processing systems, 30, 2017. L. Deng. The mnist database of handwritten digit images for machine learning research [best of the web]. IEEE signal processing magazine, 29(6):141–142, 2012. H. Xiao, K. Rasul, and R. Vollgraf. Fashion-mnist: a novel image dataset for benchmarking machine learning algorithms. arXiv preprint arXiv:1708.07747, 2017. Y. Seo and K.-s. Shin. Hierarchical convolutional neural networks for fashion image classification. Expert systems with applications, 116:328–339, 2019. L. Deng. The mnist database of handwritten digit images for machine learning research [best of the web]. IEEE signal processing magazine, 29(6):141–142, 2012. Accuracy on training epoch 12 Accuracy as a function of training epoch for all the models under consideration on the MNIST and Fashion-MNIST datasets. ML-CapsNet achieved maximum model accuracy. Accuracy on training epoch 13 Accuracy as a function of training epoch for all the models under consideration on the CIFAR-10 and CIFAR-100 datasets. ML-CapsNet achieved maximum model accuracy. Predictive Analytics 14 Level-Wise Metrics Confusion Matrix Per Hierarchy Accuracy, Precision, Recall and F1-Score per level Hierarchical