Kuzushiji-MNIST

Name: Kuzushiji-MNIST
License: CC BY-SA 4.0

Dataset Information

Modalities

Images

Languages

Japanese

License

CC BY-SA 4.0

Homepage

Official Website

Contents

Overview
Associated Benchmarks
Recent Benchmark Submissions
Research Papers

Overview

Kuzushiji-MNIST is a drop-in replacement for the MNIST dataset (28x28 grayscale, 70,000 images). Since MNIST restricts us to 10 classes, the authors chose one character to represent each of the 10 rows of Hiragana when creating Kuzushiji-MNIST. Kuzushiji is a Japanese cursive writing style.

Source: Deep Learning for Classical Japanese Literature
Image Source: https://github.com/rois-codh/kmnist

Variants: Kuzushiji-MNIST

Associated Benchmarks

This dataset is used in 2 benchmarks:

Image Classification - Metrics: Accuracy, Error, Trainable Parameters
Fine-Grained Image Classification - Metrics: Accuracy

Recent Benchmark Submissions

Task	Model	Paper	Date
Image Classification	R-ExplaiNet-26	Learning local discrete features in …	2024-10-31
Image Classification	ResNet-14	CNN Filter DB: An Empirical …	2022-03-29
Fine-Grained Image Classification	VGG-5	ProgressiveSpinalNet architecture for FC layers	2021-03-21
Image Classification	VGG-5 (Spinal FC)	SpinalNet: Deep Neural Network with …	2020-07-07
Image Classification	FWD	Multi-Complementary and Unlabeled Learning for …	2020-01-13
Image Classification	linear/flexible model	Multi-Complementary and Unlabeled Learning for …	2020-01-13
Image Classification	KerCNN	KerCNNs: biologically inspired lateral connections …	2019-10-18
Image Classification	CAMNet3	Context-Aware Multipath Networks	2019-07-26
Image Classification	Convolutional Tsetlin Machine	The Convolutional Tsetlin Machine	2019-05-23
Image Classification	Resnet-152	A Comprehensive Study of ImageNet …	2019-05-22
Image Classification	VGG8B(2x) + LocalLearning + CO	Training Neural Networks with Local …	2019-01-20
Image Classification	ResNet18 + VGG Ensemble	Deep Learning for Classical Japanese …	2018-12-03
Image Classification	Complementary-Label Learning	Complementary-Label Learning for Arbitrary Losses …	2018-10-10
Image Classification	PreActResNet-18 + Input Mixup	mixup: Beyond Empirical Risk Minimization	2017-10-25
Image Classification	PreActResNet-18	Identity Mappings in Deep Residual …	2016-03-16

Research Papers

Recent papers with results on this dataset:

Learning local discrete features in explainable-by-design convolutional neural networks (2024) -
CNN Filter DB: An Empirical Investigation of Trained Convolutional Filters (2022) -
ProgressiveSpinalNet architecture for FC layers (2021) -
SpinalNet: Deep Neural Network with Gradual Input (2020) -
Multi-Complementary and Unlabeled Learning for Arbitrary Losses and Models (2020) -
KerCNNs: biologically inspired lateral connections for classification of corrupted images (2019) -
Context-Aware Multipath Networks (2019) -
The Convolutional Tsetlin Machine (2019) -
A Comprehensive Study of ImageNet Pre-Training for Historical Document Image Analysis (2019) -
Training Neural Networks with Local Error Signals (2019) -
Deep Learning for Classical Japanese Literature (2018) -
Complementary-Label Learning for Arbitrary Losses and Models (2018) -
mixup: Beyond Empirical Risk Minimization (2017) -
Identity Mappings in Deep Residual Networks (2016) -

External Links:

Kuzushiji-MNIST

Overview edit

Associated Benchmarks

Recent Benchmark Submissions

Research Papers

Edit Dataset Information

Overview