Saved in:
Bibliographic Details
Main Authors: Wendt, Veronica, Steiner, Jacob, Yu, Byunggu, Kelly, Caleb, Kim, Justin
Format: Preprint
Published: 2024
Subjects:
Online Access:https://arxiv.org/abs/2401.07390
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866910066754453504
author Wendt, Veronica
Steiner, Jacob
Yu, Byunggu
Kelly, Caleb
Kim, Justin
author_facet Wendt, Veronica
Steiner, Jacob
Yu, Byunggu
Kelly, Caleb
Kim, Justin
contents Self-attention transformers have demonstrated accuracy for image classification with smaller data sets. However, a limitation is that tests to-date are based upon single class image detection with known representation of image populations. For instances where the input image classes may be greater than one and test sets that lack full information on representation of image populations, accuracy calculations must adapt. The Receiver Operating Characteristic (ROC) accuracy threshold can address the instances of multiclass input images. However, this approach is unsuitable in instances where image population representation is unknown. We then consider calculating accuracy using the knee method to determine threshold values on an ad-hoc basis. Results of ROC curve and knee thresholds for a multi-class data set, created from CIFAR-10 images, are discussed for multiclass image detection.
format Preprint
id arxiv_https___arxiv_org_abs_2401_07390
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Knee or ROC
Wendt, Veronica
Steiner, Jacob
Yu, Byunggu
Kelly, Caleb
Kim, Justin
Machine Learning
Computer Vision and Pattern Recognition
Self-attention transformers have demonstrated accuracy for image classification with smaller data sets. However, a limitation is that tests to-date are based upon single class image detection with known representation of image populations. For instances where the input image classes may be greater than one and test sets that lack full information on representation of image populations, accuracy calculations must adapt. The Receiver Operating Characteristic (ROC) accuracy threshold can address the instances of multiclass input images. However, this approach is unsuitable in instances where image population representation is unknown. We then consider calculating accuracy using the knee method to determine threshold values on an ad-hoc basis. Results of ROC curve and knee thresholds for a multi-class data set, created from CIFAR-10 images, are discussed for multiclass image detection.
title Knee or ROC
topic Machine Learning
Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2401.07390