Challenges in Mushroom Classification Using Image Recognition Technology

Vasiljević, M. and Vasiljević, P. and Markoski, B. and Kupusinac, A. (2018) Challenges in Mushroom Classification Using Image Recognition Technology. “St Kliment Ohridski” University - Bitola, Faculty of Information and Communication Technologies - Bitola, Republic of Macedonia, pp. 120-123. ISBN 978-9989-870-80-4

Full text not available from this repository.
Official URL: https://aiitconference.org/2018/proceedings

Abstract

In the age of technology, the competitive nature of software development could lead to the release of potentially dangerous applications. Mushroom identification applications using image recognition technology are becoming increasingly common, yet their proclaimed accuracies have not been independently ascertained. This study strives to compare the prediction accuracy of the most highly-rated mushroom identification Android applications. It was observed that the accuracy in classifying the correct mushroom species ranged from twenty to fifty percent, depending on the number of images per species provided to the application. This was far below the accuracies declared by the developers. The danger in this discrepancy is that mushroom pickers who are amateur taxonomists could mistake a poisonous species for an edible one, leading to deadly consequences. It is therefore recommended that app developers include scientific-based taxonomic methods during software development and also increase the number of images available for image classification since this greatly improves accuracy.

Item Type: Book
Subjects: Scientific Fields (Frascati) > Engineering and Technology > Electrical engineering, electronic engineering,information engineering
Divisions: Faculty of Information and Communication Technologies
Depositing User: Mrs Natasha Tabakovska
Date Deposited: 03 Mar 2019 16:50
Last Modified: 13 May 2019 08:51
URI: http://eprints.uklo.edu.mk/id/eprint/1790

Actions (login required)

View Item View Item