Ajou University repository

A visualization system for performance analysis of image classification modelsoa mark
Citations

SCOPUS

0

Citation Export

Publication Year
2020-01-26
Journal
IS and T International Symposium on Electronic Imaging Science and Technology
Publisher
Society for Imaging Science and Technology
Citation
IS and T International Symposium on Electronic Imaging Science and Technology, Vol.2020 No.1
Mesh Keyword
Classification modelsComparison and analysisMachine learning modelsModel interpretationsModel visualizationPerformance analysisVisual analytics systemsVisualization system
All Science Classification Codes (ASJC)
Computer Graphics and Computer-Aided DesignComputer Science ApplicationsHuman-Computer InteractionSoftwareElectrical and Electronic EngineeringAtomic and Molecular Physics, and Optics
Abstract
Developing machine learning models for image classification problems involves various tasks such as model selection, layer design, and hyperparameter tuning for improving the model performance. However, regarding deep learning models, insufficient model interpretability renders it infeasible to understand how they make predictions. To facilitate model interpretation, performance analysis at the class and instance levels with model visualization is essential. We herein present an interactive visual analytics system to provide a wide range of performance evaluations of different machine learning models for image classification. The proposed system aims to overcome challenges by providing visual performance analysis at different levels and visualizing misclassification instances. The system which comprises five views - ranking, projection, matrix, and instance list views, enables the comparison and analysis different models through user interaction. Several use cases of the proposed system are described and the application of the system based on MNIST data is explained. Our demo app is available at https://chanhee13p.github.io/VisMlic/.
ISSN
2470-1173
Language
eng
URI
https://aurora.ajou.ac.kr/handle/2018.oak/36615
https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85095567161&origin=inward
DOI
https://doi.org/10.2352/issn.2470-1173.2020.1.vda-375
Type
Conference
Funding
This work was performed under the auspices of the U.S. Department of Energy by Lawrence Livermore National Laboratory under Contract DE-AC52-07NA27344.
Show full item record

Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.

Related Researcher

Lee, KyungWon Image
Lee, KyungWon이경원
Department of Digital Media
Read More

Total Views & Downloads

File Download