Symmetric Positive Definite Representation

Symmetric Positive Definite (SPD) matrix based representation is widely used in many visual recognition tasks. This archive is an effort for keeping the track of this thriving and novel field of research in computer vision.

Introduction


Recently, symmetric positive definitive (SPD) matrix-based visual representation methods have shown promising performance in various applications such as fine-grained image classification, person re-identification and ImageNet classification. This page keeps track of the recent advances in SPD matrix-based visual representation methods. Kindly refer to the contact section if you have any queries or sugesstions, and interested to add your method in the listings.

Workshops and tutorials on SPD representation


Tutorial on Higher-order Statistical Modeling based Deep Convolutional Neural Networks

  • by Peihua Li, Wangmeng Zuo, Qilong Wang
  • at Chinese Conference on Pattern Recognition and Computer Vision (PRCV), 2018
Tutorial Page

Tutorial on Second- and Higher-order Representations in Computer Vision

  • by Piotr Koniusz, Mehrtash Harandi, Lei Wang, Ruiping Wang
  • at International Conference on Computer Vision (ICCV), 2019
Tutorial Page

Selective SPD representation methods


Bilinear CNNs for Fine-grained Visual Recognition

  • by Tsung-Yu Lin, Aruni RoyChowdhury, Subhransu Maji
  • at International Conference on Computer Vision (ICCV), 2015
Paper Link Source Code Project Page

Compact Bilinear Pooling

  • by Yang Gao, Oscar Beijbom, Ning Zhang, Trevor Darrell
  • at International Conference on Computer Vision and Pattern Recognition (CVPR), 2016
Paper Link Source Code

Kernel Pooling for Convolutional Neural Networks

  • by Yin Cui, Feng Zhou, Jiang Wang, Xiao Liu, Yuanqing Lin, Serge Belongie
  • at International Conference on Computer Vision and Pattern Recognition (CVPR), 2017
Paper Link Project Page

Low-rank Bilinear Pooling for Fine-Grained Classification

  • by Shu Kong, Charless Fowlkes
  • at International Conference on Computer Vision and Pattern Recognition (CVPR), 2017
Paper Link Source Code Project Page

Is Second-order Information Helpful for Large-scale Visual Recognition?

  • by Peihua Li, Jiangtao Xie, Qilong Wang, Wangmeng Zuo
  • at International Conference on Computer Vision (ICCV), 2017
Paper Link Source Code Project Page

Improved Bilinear Pooling with CNNs

  • by Tsung-Yu Lin, Subhransu Maji
  • at British Machine Vision Conference (BMVC), 2017
Paper Link Source Code Project Page

Factorized Bilinear Models for Image Recognition

  • by Yanghao Li, Naiyan Wang, Jiaying Liu, Xiaodi Hou
  • at International Conference on Computer Vision (ICCV), 2017
Paper Link Source Code Project Page

G2DeNet: Global Gaussian Distribution Embedding Network and Its Application to Visual Recognition

  • by Qilong Wang, Peihua Li, Lei Zhang
  • at International Conference on Computer Vision and Pattern Recognition (CVPR), 2017
Paper Link Source Code

Towards Faster Training of Global Covariance Pooling Networks by Iterative Matrix Square Root Normalization

  • by Peihua Li, Jiangtao Xie, Qilong Wang, Zilin Gao
  • at International Conference on Computer Vision and Pattern Recognition (CVPR), 2018
Paper Link Source Code Project Page

Hierarchical Bilinear Pooling for Fine-Grained Visual Recognition

  • by Chaojian Yu, Xinyi Zhao, Qi Zheng, Peng Zhang, Xinge You
  • at European Conference on Computer Vision (ECCV), 2018
Paper Link Source Code

Statistically-motivated Second-order Pooling

  • by Kaicheng Yu, Mathieu Salzmann
  • at European Conference on Computer Vision (ECCV), 2018
Paper Link Source Code

DeepKSPD: learning kernel-matrix-based SPD representation for fine-grained image recognition

  • by Melih Engin, Lei Wang, Luping Zhou, Xinwang Liu
  • at European Conference on Computer Vision (ECCV), 2018
Paper Link

Second-order Democratic Aggregation

  • by Tsung-Yu Lin, Subhransu Maji, Piotr Koniusz
  • at European Conference on Computer Vision (ECCV), 2018
Paper Link Source Code Project Page

Global Gated Mixture of Second-order Pooling for Improving Deep Convolutional Neural Networks

  • by Qilong Wang, Zilin Gao, Jiangtao Xie, Wangmeng Zuo and Peihua Li
  • at Advances in Neural Information Processing Systems (NIPS), 2018
Paper Link Source Code

Local Temporal Bilinear Pooling for Fine-Grained Action Parsing

  • by Yan Zhang, Siyu Tang, Krikamol Muandet, Christian Jarvers, Heiko Neumann
  • at International Conference on Computer Vision and Pattern Recognition (CVPR), 2019
Paper Link Source Code

Global Second-order Pooling Convolutional Networks

  • by Zilin Gao, Jiangtao Xie, Qilong Wang, Peihua Li
  • at International Conference on Computer Vision and Pattern Recognition (CVPR), 2019
Paper Link

Deep Global Generalized Gaussian Networks

  • by Qilong Wang, Peihua Li, Qinghua Hu, Pengfei Zhu, Wangmeng Zuo.
  • at International Conference on Computer Vision and Pattern Recognition (CVPR), 2019
Paper Link

Second-order Attention Network for Single Image Super-Resolution

  • by Tao Dai, Jianrui Cai, Yongbing Zhang, Shu-Tao Xia, Lei Zhang
  • at International Conference on Computer Vision and Pattern Recognition (CVPR), 2019
Paper Link Source Code

Low-Rank Pairwise Alignment Bilinear Network For Few-Shot Fine-Grained Image Classification

  • by Huaxi Huang, Junjie Zhang, Jian Zhang, Jingsong Xu, Qiang Wu
  • at arXiv:1908.01313v1, 2019
Paper Link

Compact Approximation for Polynomial of Covariance Feature

  • by Yusuke Mukuta, Tatsuaki Machida, Tatsuya Harada
  • at arXiv:1906.01851, 2019.
Paper Link

Learning Neural Bag-of-Matrix-Summarization with Riemannian Network

  • by Hong Liu, Jie Li, Rongrong Ji, Yongjian Wu
  • at AAAI Conference on Artificial Intelligence (AAAI), 2019
Paper Link

Fine-Grained Classification via Hierarchical Bilinear Pooling With Aggregated Slack Mask

  • by Min Tan, Guijun Wang, Jian Zhou, Zhiyou Peng, Meilian Zheng
  • at IEEE Access, Volume 7, pages 117944-117953, 2019
Paper Link Source Code Project Page

Learning Deep Bilinear Transformation for Fine-grained Image Representation

  • by Heliang Zheng, Jianlong Fu, Zheng-Jun Zha, Jiebo Luo
  • at Conference on Neural Information Processing Systems (NIPS), 2019
Paper Link Source Code Project Page

Applications of SPD representation


Fundamental Research

Higher-Order Occurrence Pooling for Bags-of-Words: Visual Concept Detection

  • by Piotr Koniusz, Fei Yan, Philippe-Henri Gosselin, Krystian Mikolajczyk
  • at IEEE Transactions on Pattern Analysis and Machine Intelligence, Volume 39, No. 2, Pages 313-326, Feb. 2017
Paper Link

A Deeper Look at Power Normalizations

  • by Piotr Koniusz, Hongguang Zhang, Fatih Porikli
  • at International Conference on Computer Vision and Pattern Recognition (CVPR), 2018
Paper Link

A Robust Distance Measure for Similarity-Based Classification on the SPD Manifold

  • by Zhi Gao, Yuwei Wu, Mehrtash Tafazzoli Harandi, Yunde Jia
  • at IEEE Transactions on Neural Networks and Learning Systems, 2019
Paper Link

What Deep CNNs Benefit from Global Covariance Pooling: An Optimization Perspective

  • by Qilong Wang, Liyong Zhang, Banggu Wu, Dongwei Ren, P. Li, W. Zuo, Q. Hu
  • at International Conference on Computer Vision and Pattern Recognition (CVPR), 2020
Paper Link Source Code

Revisiting Bilinear Pooling: A Coding Perspective

  • by Z. Gao, Yuwei Wu, Xiaoxun Zhang, Jindou Dai, Y. Jia, M. Harandi
  • at AAAI Conference on Artificial Intelligence, 2020
Paper Link

Power Normalizations in Fine-grained Image, Few-shot Image and Graph Classification

  • by Piotr Koniusz and Hongguang Zhang
  • at IEEE Transactions on Pattern Analysis and Machine Intelligence, 2020
Paper Link

Fine-grained/Texture image classification

Bilinear CNNs for Fine-grained Visual Recognition

  • by Tsung-Yu Lin, Aruni RoyChowdhury, Subhransu Maji
  • at International Conference on Computer Vision (ICCV), 2015
Paper Link Source Code Project Page

Compact Bilinear Pooling

  • by Yang Gao, Oscar Beijbom, Ning Zhang, Trevor Darrell
  • at International Conference on Computer Vision and Pattern Recognition (CVPR), 2016
Paper Link Source Code

Sparse Coding for Third-order Super-symmetric Tensor Descriptors with Application to Texture Recognition

  • by Piotr Koniusz, Anoop Cherian
  • at International Conference on Computer Vision and Pattern Recognition (CVPR), 2016
Paper Link

Kernel Pooling for Convolutional Neural Networks

  • by Yin Cui, Feng Zhou, Jiang Wang, Xiao Liu, Yuanqing Lin, Serge Belongie
  • at International Conference on Computer Vision and Pattern Recognition (CVPR), 2017
Paper Link Project Page

Low-rank Bilinear Pooling for Fine-Grained Classification

  • by Shu Kong, Charless Fowlkes
  • at International Conference on Computer Vision and Pattern Recognition (CVPR), 2017
Paper Link Source Code Project Page

Improved Bilinear Pooling with CNNs

  • by Tsung-Yu Lin, Subhransu Maji
  • at British Machine Vision Conference (BMVC), 2017
Paper Link Source Code Project Page

Factorized Bilinear Models for Image Recognition

  • by Yanghao Li, Naiyan Wang, Jiaying Liu, Xiaodi Hou
  • at International Conference on Computer Vision (ICCV), 2017
Paper Link Source Code Project Page

G2DeNet: Global Gaussian Distribution Embedding Network and Its Application to Visual Recognition

  • by Qilong Wang, Peihua Li, Lei Zhang
  • at International Conference on Computer Vision and Pattern Recognition (CVPR), 2017
Paper Link Source Code

Where to Focus: Deep Attention-based Spatially Recurrent Bilinear Networks forFine-Grained Visual Recognition

  • by Lin Wu, Yang Wang
  • at arXiv:1709.05769, 2017
Paper Link Source Code

Towards Faster Training of Global Covariance Pooling Networks by Iterative Matrix Square Root Normalization

  • by Peihua Li, Jiangtao Xie, Qilong Wang, Zilin Gao
  • at International Conference on Computer Vision and Pattern Recognition (CVPR), 2018
Paper Link Source Code Project Page

Hierarchical Bilinear Pooling for Fine-Grained Visual Recognition

  • by Chaojian Yu, Xinyi Zhao, Qi Zheng, Peng Zhang, Xinge You
  • at European Conference on Computer Vision (ECCV), 2018
Paper Link Source Code

Grassmann Pooling as Compact Homogeneous Bilinear Pooling for Fine-Grained Visual Classification

  • by Xing Wei, Yue Zhang, Yihong Gong, Jiawei Zhang, Nanning Zheng
  • at International Conference on Computer Vision and Pattern Recognition (CVPR), 2018
Paper Link

Statistically-motivated Second-order Pooling

  • by Kaicheng Yu, Mathieu Salzmann
  • at European Conference on Computer Vision (ECCV), 2018
Paper Link Source Code

DeepKSPD: learning kernel-matrix-based SPD representation for fine-grained image recognition

  • by Melih Engin, Lei Wang, Luping Zhou, Xinwang Liu
  • at European Conference on Computer Vision (ECCV), 2018
Paper Link

Second-order Democratic Aggregation

  • by Tsung-Yu Lin, Subhransu Maji, Piotr Koniusz
  • at European Conference on Computer Vision (ECCV), 2018
Paper Link Source Code Project Page

Fine-Grained Image Classification With Gaussian Mixture Layer

  • by Jingyun Liang, Jinlin Guo, Xin Liu, Songyang Lao
  • at IEEE Access, Volume 6, pages 53356-53367, 2018
Paper Link

Learning a Robust Representation via a Deep Network on Symmetric Positive Definite Manifolds

  • by Zhi Gao, Yuwei Wu, Xingyuan Bu, Yunde Jia
  • at Pattern Recognition, Volume 92, Pages 1-12, 2019
Paper Link

Compact Approximation for Polynomial of Covariance Feature

  • by Yusuke Mukuta, Tatsuaki Machida, Tatsuya Harada
  • at arXiv:1906.01851, 2019.
Paper Link

Fine-Grained Classification via Hierarchical Bilinear Pooling With Aggregated Slack Mask

  • by Min Tan, Guijun Wang, Jian Zhou, Zhiyou Peng, Meilian Zheng
  • at IEEE Access, Volume 7, pages 117944-117953, 2019
Paper Link Source Code

Learning Deep Bilinear Transformation for Fine-grained Image Representation

  • by Heliang Zheng, Jianlong Fu, Zheng-Jun Zha, Jiebo Luo
  • at Conference on Neural Information Processing Systems (NIPS), 2019
Paper Link Source Code Project Page

ReDro: Efficiently Learning Large-Sized SPD Visual Representation

  • by S. Rahman, Lei Wang, Changming Sun, L. Zhou
  • at European Conference on Computer Vision (ECCV), 2020
Paper Link

ImageNet/Generic image classification

Is Second-order Information Helpful for Large-scale Visual Recognition?

  • by Peihua Li, Jiangtao Xie, Qilong Wang, Wangmeng Zuo
  • at International Conference on Computer Vision (ICCV), 2017
Paper Link Source Code Project Page

Towards Faster Training of Global Covariance Pooling Networks by Iterative Matrix Square Root Normalization

  • by Peihua Li, Jiangtao Xie, Qilong Wang, Zilin Gao
  • at International Conference on Computer Vision and Pattern Recognition (CVPR), 2018
Paper Link Source Code Project Page

Global Gated Mixture of Second-order Pooling for Improving Deep Convolutional Neural Networks

  • by Qilong Wang, Zilin Gao, Jiangtao Xie, Wangmeng Zuo and Peihua Li
  • at Advances in Neural Information Processing Systems (NIPS), 2018
Paper Link Source Code

Global Second-order Pooling Convolutional Networks

  • by Zilin Gao, Jiangtao Xie, Qilong Wang, Peihua Li
  • at International Conference on Computer Vision and Pattern Recognition (CVPR), 2019
Paper Link Source Code

Deep Global Generalized Gaussian Networks

  • by Qilong Wang, Peihua Li, Qinghua Hu, Pengfei Zhu, Wangmeng Zuo
  • at International Conference on Computer Vision and Pattern Recognition (CVPR), 2019
Paper Link

Action parsing/Temporal modelling

Tensor Representations via Kernel Linearization for Action Recognition from 3D Skeletons

  • by Piotr Koniusz, Anoop Cherian, Fatih Porikli
  • at European Conference on Computer Vision (ECCV), 2016
Paper Link

Higher-Order Pooling of CNN Features via Kernel Linearization for Action Recognition

  • by Anoop Cherian, Piotr Koniusz, Stephen Gould
  • at Winter Conference on Applications of Computer Vision (WACV), 2017
Paper Link

Local Temporal Bilinear Pooling for Fine-Grained Action Parsing

  • by Yan Zhang, Siyu Tang, Krikamol Muandet, Christian Jarvers, Heiko Neumann
  • at International Conference on Computer Vision and Pattern Recognition (CVPR), 2019
Paper Link Source Code

Approximated Bilinear Modules for Temporal Modeling

  • by Xinqi Zhu, Chang Xu, Langwen Hui, Cewu Lu, Dacheng Tao
  • at International Conference on Computer Vision (ICCV), 2019
Paper Link Source Code

Tensor Representations for Action Recognition

  • by Piotr Koniusz, Lei Wang, Anoop Cherian
  • at IEEE Transactions on Pattern Analysis and Machine Intelligence, 2020
Paper Link

Person Re-indentification

Second-Order Non-Local Attention Networks for Person Re-Identification

  • by Bryan (Ning) Xia, Yuan Gong, Yizhe Zhang, Christian Poellabauer
  • at International Conference on Computer Vision (ICCV), 2019
Paper Link

Mixed High-Order Attention Network for Person Re-Identification

  • by Binghui Chen, Weihong Deng, Jiani Hu
  • at International Conference on Computer Vision (ICCV), 2019
Paper Link Source Code

High-Order Information Matters: Learning Relation and Topology for Occluded Person Re-Identification

  • by G. Wang, S. Yang, Huanyu Liu, Z. Wang, Yang Yang, Shuliang Wang, Gang Yu, Erjin Zhou, J. Sun
  • at International Conference on Computer Vision and Pattern Recognition (CVPR), 2020
Paper Link Source Code

Second-order Camera-aware Color Transformation for Cross-domain Person Re-identification

  • by Wangmeng Xiang, Hongwei Yong, Jianqiang Huang, Xian-Sheng Hua, Lei Zhang
  • at Asian Conference on Computer Vision (ACCV), 2020
Paper Link

Domain Adaptation

Domain Adaptation by Mixture of Alignments of Second- or Higher-Order Scatter Tensors

  • by Piotr Koniusz, Yusuf Tas, Fatih Porikli
  • at International Conference on Computer Vision and Pattern Recognition (CVPR), 2016
Paper Link

Museum Exhibit Identification Challenge for Domain Adaptation and Beyond

  • by Piotr Koniusz, Yusuf Tas, Hongguang Zhang, Mehrtash Harandi, Fatih Porikli, Rui Zhang
  • at European Conference on Computer Vision (ECCV), 2018
Paper Link

Domain Adaptation Using Riemannian Geometry of SPD Matrices

  • by Gal Maman, Or Yair, Danny Eytan, Ronen Talmon
  • at International Conference on Acoustics, Speech and Signal Processing (ICASSP), 2019
Paper Link

Few-shot Learning

Power Normalizing Second-Order Similarity Network for Few-Shot Learning

  • by Hongguang Zhang, Piotr Koniusz
  • at Winter Conference on Applications of Computer Vision (WACV), 2018
Paper Link

Few-Shot Learning via Saliency-guided Hallucination of Samples

  • by Hongguang Zhang, Jing Zhang, Piotr Koniusz
  • at International Conference on Computer Vision and Pattern Recognition (CVPR), 2019
Paper Link

Few-Shot Object Detection by Second-order Pooling

  • by Shan Zhang, Dawei Luo, Lei Wang
  • at Asian Conference on Computer Vision (ACCV), 2020
Paper Link

Few-shot Action Recognition with Permutation-invariant Attention

  • by Hongguang Zhang, Li Zhang, Xiaojuan Qi, Hongdong Li, Philip Torr, Piotr Koniusz
  • at European Conference on Computer Vision (ECCV), 2020
Paper Link

Adaptive Subspaces for Few-Shot Learning

  • by Christian Simon, Piotr Koniusz, Richard Nock, Mehrtash Harandi
  • at International Conference on Computer Vision and Pattern Recognition (CVPR), 2020
Paper Link

Others

RAID-G: Robust Estimation of Approximate Infinite Dimensional Gaussian with Application to Material Recognition

  • by Qilong Wang, Peihua Li1, Wangmeng Zuo, Lei Zhang
  • at International Conference on Computer Vision and Pattern Recognition (CVPR), 2016
Paper Link Source Code

A Neural Network Based on SPD Manifold Learning for Skeleton-Based Hand Gesture Recognition

  • by Xuan Son Nguyen, Luc Brun, Olivier Lezoray, Sebastien Bougleux
  • at International Conference on Computer Vision and Pattern Recognition (CVPR), 2019
Paper Link Source Code

Second-Order Attention Network for Single Image Super-Resolution

  • by Tao Dai, Jianrui Cai, Yongbing Zhang, Shu-Tao Xia, Lei Zhang
  • at International Conference on Computer Vision and Pattern Recognition (CVPR), 2019
Paper Link Source Code

Bilinear Attention Networks for Person Retrieval

  • by Pengfei Fang, Jieming Zhou, Soumava Kumar Roy, Lars Petersson, Mehrtash Harandi
  • at International Conference on Computer Vision (ICCV), 2019
Paper Link

Factorized Higher-Order CNNs with an Application to Spatio-Temporal Emotion Estimation

  • by Jean Kossaifi, Antoine Toisoul, Adrian Bulat, Yannis Panagakis, Timothy M. Hospedales, M. Pantic
  • at International Conference on Computer Vision and Pattern Recognition (CVPR), 2020
Paper Link

DOA-GAN: Dual-Order Attentive Generative Adversarial Network for Image Copy-move Forgery Detection and Localization

  • by Ashraful Islam, Chengjiang Long, A. Basharat, A. Hoogs
  • at International Conference on Computer Vision and Pattern Recognition (CVPR), 2020
Paper Link

Visual-Semantic Matching by Exploring High-Order Attention and Distraction

  • by Yongzhi Li, D. Zhang, Y. Mu
  • at International Conference on Computer Vision and Pattern Recognition (CVPR), 2020
Paper Link

Second Order enhanced Multi-glimpse Attention in Visual Question Answering

  • by Qiang Sun, Binghui Xie, Yanwei Fu
  • at Asian Conference on Computer Vision (ACCV), 2020
Paper Link

SOFA-Net: Second-Order and First-order Attention Network for Crowd Counting

  • by Haoran Duan, Shidong Wang, Yu Guan
  • at British Machine Vision Conference (BMVC), 2020
Paper Link

BARNet: Bilinear Attention Network with Adaptive Receptive Field for Surgical Instrument Segmentation

  • by Zhen-Liang Ni, Guibin Bian, G. Wang, Xiao-Hu Zhou, Zengguang Hou, Xiaoliang Xie, Zhuguo Li, Yu-Han Wang
  • at International Joint Conference on Artificial Intelligence (IJCAI), 2020
Paper Link

Non-Local Neural Networks with Grouped Bilinear Attentional Transforms

  • by L. Chi, Ze-Huan Yuan, Y. Mu, C. Wang
  • at International Conference on Computer Vision and Pattern Recognition (CVPR), 2020
Paper Link

Analysis of existing methods and their performance


Plese feel free to use sorting options and search box for efficient analysis of the existing SPD representation based methods and their results.

Summary of method Performace across datasets (in %)
Year Method Backbone Model Classifier Input Size DA Optimizer Feature Size CUB Airplane Cars MIT DTD Food-101 ImageNet
2015 Bilinear CNN VGG-M/VGG-16 SVM 448X448 Flip SGD 262k 84.1 84.1 91.3 -- -- -- --
2016 RAID-G VGG19 SVM 224X224 -- -- -- 84.0 -- -- -- 76.4 -- --
2016 Compact Bilinear VGGM/VGG16 SVM 448X448 -- SGD 10k 84.0 -- -- 73.4 67.7 -- --
2017 Kernel Pooling VGG16/ResNet50 SVM 448X448 Flip, Crop SGD 12.8k/14.3k 86.2 86.9 92.4 -- -- 85.5 --
2017 Low-Rank Bilinear VGG16/ResNet50 SVM 448X448 Flip, Crop SGD 12.8k/14.3k 84.2 87.3 90.9 -- 65.8 -- --
2017 MPN-COV VGG16/ResNet50 SVM 448X448 Jitter SGD 32k -- -- -- -- -- -- 78.8
2017 Improved BCNN VGGM/VGG16 SVM 448X448 Flip SGD 262k 85.8 88.5 92.0 -- -- -- --
2017 Factorized bilinear ResNet50 SVM 112X112 Flip, Crop SGD 10M 82.9 -- -- -- 67.8 -- 76.0
2017 G2DeNet VGG16 SVM 448X448 Flip SGD 131k 87.1 89.0 92.5 -- -- -- --
2017 Recurrent Bilinear VGG16 SVM 448X448 Flip, Shift SGD 262k 89.7 88.4 93.4 -- -- -- --
2018 iSQRT-COV AlexNet/ResNet101 Softmax 224X224 Flip, Crop, Jitter SGD 32k 88.7 91.4 93.3 -- -- -- 78.79
2018 Grassmann Pooling VGG16 Softmax 448X448 Flip SGD 4k 85.8 89.8 92.8 -- -- 85.7 --
2018 Hierarchical Bilinear VGG16 Softmax 448X448 Flip, Crop SGD 8k 87.1 90.3 93.7 -- -- -- --
2018 SMSO Pooling VGG16/ResNet50 Softmax 448X448 Flip SGD 2k 85.8 -- -- 79.7 72.5 -- --
2018 DeepKSPD VGG16 Softmax 448X448 Flip Adam 262k 86.5 91.5 93.2 81.0 -- -- --
2018 SoDA VGG16/ResNet50 SVM 448X448 Flip SGD 8k/262k 85.9 87.6 91.7 84.3 76.2 -- --
2018 GM-SOP ResNet18 Softmax 64X64 Flip SGD 8k -- -- -- -- -- -- 67.7*
2018 GMNet VGG16/VGG19 Softmax random Flip SGD 400k 86.3 90.5 93.5 -- -- -- --
2020 SOP+SC+SigmE AlexNet/ResNet50 Softmax 224/336/448 Flip Adam -- -- -- -- 86.3 -- 87.5 --
2019 SPD aggregation Network VGG16 Softmax 224X224 -- SGD 262k 72.4 77.8 -- -- 68.9 -- --
2019 Global SoP ResNet50 Softmax 224X224 Flip SGD 2k/32k -- -- -- -- -- -- 78.9*
2019 3G-Net ResNet50/ResNet101 Softmax 224X224 Flip SGD 32k -- -- -- -- -- -- 79.1*
2019 iPCCP-Net ResNet50/ResNet101 Softmax 448X448 Flip SGD 8k 88.4 91.6 94.1 -- -- -- --
2019 HBPASM ResNet-34 Softmax 448X448 Flip SGD 24k 86.8 91.3 93.8 -- -- -- --
2019 DBTNet ResNet50/ResNet101 Softmax 448X448 -- SGD 2k 88.1 91.6 94.5 -- -- -- --
2020 ReDro VGG16/ResNet50 Softmax 448X448 Flip Adam 262k 84.3 89.2 92.2 84.0 -- -- --
2020 SOP+SC+Spec. MaxExp(F) AlexNet/ResNet50 Softmax 224/336/448 Flip Adam -- -- -- -- 86.8 -- 88.4 77.95

* - Produced with ImageNet-1k dataset while the others use ImageNet-2012 dataset.


Acknowledgements


This website uses images and code links that are shared by the original authors. Bootstarp 4 CSS library and jQuery 3.x were used for the development of this website.

Contact


For any queries and sugesstions you may directly contact at sr801@uowmail.edu.au or leiw@uow.edu.au. Also, please do not hesitate to contact us if you would like to add your method into the listing of SPD Archive.