DenseNet Models for Tiny ImageNet Classification

Zoheb Abai, Nishad Rajmalwar
4/23/2019
cs.CV

Official code: ZohebAbai/Tiny-ImageNet-Challenge (18 stars, MIT, last commit 2020-06-02).

Abstract

In this paper, we present two image classification models on the Tiny ImageNet dataset. We built two very different networks from scratch based on the idea of Densely Connected Convolution Networks. The architecture of the networks is designed based on the image resolution of this specific dataset and by calculating the Receptive Field of the convolution layers. We also used some non-conventional techniques related to image augmentation and Cyclical Learning Rate to improve the accuracy of our models. The networks are trained under high constraints and low computation resources. We aimed to achieve top-1 validation accuracy of 60%; the results and error analysis are also presented.

View on arXivView PDF

Code Implementations(1)

ZohebAbai/Tiny-ImageNet-ChallengeOfficial100%
MIT

This repository contains the jupyter notebooks for the custom-built DenseNet Model build on Tiny ImageNet dataset

185Jupyter NotebookApr 16, 20196 years agoMIT
cyclical-learning-ratesdensenetdensenet-modelsimage-augmentationkeras-tensorflow+3 more

Similar Papers

ResMLP: Feedforward networks for image classification with data-efficient training

85% match2 repos36.9k2021

FCC-GAN: A Fully Connected and Convolutional Net Architecture for GANs

84% match1 repo52019

LR-Net: A Block-based Convolutional Neural Network for Low-Resolution Image Classification

84% match1 repo192022

SoftReMish: A Novel Activation Function for Enhanced Convolutional Neural Networks for Visual Recognition Performance

84% match2025

Aggregated Residual Transformations for Deep Neural Networks

84% match1 repo1.9k2016

MBInception: A new Multi-Block Inception Model for Enhancing Image Processing Efficiency

84% match2024

Cite this paper

@article{abai2019densenet,
  title  = {DenseNet Models for Tiny ImageNet Classification},
  author = {Zoheb Abai and Nishad Rajmalwar},
  year   = {2019},
  eprint = {1904.10429},
  archivePrefix = {arXiv},
  url    = {https://arxiv.org/abs/1904.10429}
}

Add a badge to your README

Shows whether official code is linked for this paper.

Code of Paper badge
[![Code of Paper](https://codeofpaper.com/badge/1904.10429.svg)](https://codeofpaper.com/paper/1904.10429)

Discussion