Improving Robustness Using MixUp and CutMix Augmentation for Corn Leaf Diseases Classification based on ConvMixer Architecture

Dublin Core

Title

Improving Robustness Using MixUp and CutMix Augmentation for Corn Leaf Diseases Classification based on ConvMixer Architecture

Subject

ConvMixer; corn leaf diseases; CutMix; data augmentation; MixUp; robustness.

Description

Corn leaf diseases such as blight spot, gray leaf spot, and common rust still lurk in corn fields.Thisproblemmustbe solved to help corn farmers.The ConvMixer model,consisting of a patch embedding layer,is a new model with a simple structure. When training a modelwithConvMixer, improvisation is an important part that needs to befurther explored to achieve better accuracy. By using advanced data augmentation techniques such as MixUp and CutMix, the robustness ofConvMixermodel can be well achievedfor corn leaf diseasesclassification.We describe experimental evidence in thisarticle usingprecision, recall, accuracy score, and F1score as performance metrics. As a result,it turned out thatthe training model with thedata set without extension on the ConvMixer model achieved an accuracy of 0.9812,but this could still be improved. In fact, when we used the MixUp and CutMixaugmentation, the training model resultsincreasedsignificantlyto 0.9925 and 0.9932,respectively

Creator

Li-Hua Li & Radius Tanone*

Source

https://journals.itb.ac.id/index.php/jictra/article/view/18894/6295

Publisher

Department of Information Management, Chaoyang University of Technology,

Date

2023

Contributor

Fajar bagus W

Format

PDF

Language

English

Type

Text

Files

Citation

Li-Hua Li & Radius Tanone*, “Improving Robustness Using MixUp and CutMix Augmentation for Corn Leaf Diseases Classification based on ConvMixer Architecture,” Repository Horizon University Indonesia, accessed March 15, 2025, https://repository.horizon.ac.id/items/show/7044.