System for Automatic Visual Fungi Recognition

Kód výsledku v IS VaVaI

<a href="https://www.isvavai.cz/riv?ss=detail&h=RIV%2F49777513%3A23520%2F20%3A43960895" target="_blank" >RIV/49777513:23520/20:43960895 - isvavai.cz</a>

Výsledek na webu

<a href="http://www.kky.zcu.cz/cs/sw/fungi" target="_blank" >http://www.kky.zcu.cz/cs/sw/fungi</a>

DOI - Digital Object Identifier

—

Jazyk výsledku

angličtina

Název v původním jazyce

System for Automatic Visual Fungi Recognition

Popis výsledku v původním jazyce

System for automatic Fungi species recognition was made in conjunction with the Danish Mycological Society and Google. We created a Machine Learning driven system for recognition of the 1394 fungi species allowing users to automatically identify observed specimens while providing valuable data to mycologists and computer vision researchers. Following the advances in deep learning for fine-grained image categorization, our approach is based on Convolutional Neural Networks. The system is wrapped up around our winning submission to the FGVCx Fungi Classification Kaggle competition organized in connection with the CVPR2018 Fine-Grained Visual Categorization (FGVC) workshop. The neural networks were fine-tuned with te Tensorflow-Slim framework, including advanced techniques to achieve a balance between inference speed and classification performance. For the Danish Fungal Atlas, we provide a REST API used as a species proposing tool after an image is submitted. Additionally, we publish two optimized neural network models (Inception-V4 and MobileNet-V3) targeted to be run on edge devices, e.g., mobile phones or web browsers, and web application for real-time fungi recognition. The Web Application integrates the MobileNet-V3 directly in the web browser and allows users to identify Fungal species on a given image automatically. The created system has a vast potential to increase human involvement with nature by providing a real-time electronic identification tool that can intuitively support learning, much like children learn from their parents by asking simple and naive questions addressed thoroughly. By linking the system to an existing mycological platform involving validation by the community, as is the case in the Danish Fungal Atlas, a supervised machine learning system with a human in the loop is created. Additionally, the image recognition system&apos;s usage makes community-based fungi observation identification easier: from the first 592 approved annotations, 89% were based on the top-2 predictions of our model. The software was created by researchers from the Faculty of Applied Sciences - University of West Bohemia (UWB) and the Faculty of Electrical Engineering - Czech Technical University (CTU) in Prague and supported by UWB grant SGS-2019-027 and CTU grants SGS17/185/OHK3/3T/13 and CZ.02.1.01/0.0/0.0/16019/0000765.

Název v anglickém jazyce

System for Automatic Visual Fungi Recognition

Popis výsledku anglicky

System for automatic Fungi species recognition was made in conjunction with the Danish Mycological Society and Google. We created a Machine Learning driven system for recognition of the 1394 fungi species allowing users to automatically identify observed specimens while providing valuable data to mycologists and computer vision researchers. Following the advances in deep learning for fine-grained image categorization, our approach is based on Convolutional Neural Networks. The system is wrapped up around our winning submission to the FGVCx Fungi Classification Kaggle competition organized in connection with the CVPR2018 Fine-Grained Visual Categorization (FGVC) workshop. The neural networks were fine-tuned with te Tensorflow-Slim framework, including advanced techniques to achieve a balance between inference speed and classification performance. For the Danish Fungal Atlas, we provide a REST API used as a species proposing tool after an image is submitted. Additionally, we publish two optimized neural network models (Inception-V4 and MobileNet-V3) targeted to be run on edge devices, e.g., mobile phones or web browsers, and web application for real-time fungi recognition. The Web Application integrates the MobileNet-V3 directly in the web browser and allows users to identify Fungal species on a given image automatically. The created system has a vast potential to increase human involvement with nature by providing a real-time electronic identification tool that can intuitively support learning, much like children learn from their parents by asking simple and naive questions addressed thoroughly. By linking the system to an existing mycological platform involving validation by the community, as is the case in the Danish Fungal Atlas, a supervised machine learning system with a human in the loop is created. Additionally, the image recognition system&apos;s usage makes community-based fungi observation identification easier: from the first 592 approved annotations, 89% were based on the top-2 predictions of our model. The software was created by researchers from the Faculty of Applied Sciences - University of West Bohemia (UWB) and the Faculty of Electrical Engineering - Czech Technical University (CTU) in Prague and supported by UWB grant SGS-2019-027 and CTU grants SGS17/185/OHK3/3T/13 and CZ.02.1.01/0.0/0.0/16019/0000765.

Druh

R - Software

CEP obor

—

OECD FORD obor

10201 - Computer sciences, information science, bioinformathics (hardware development to be 2.2, social aspect to be 5.8)

Projekt

<a href="/cs/project/LO1506" target="_blank" >LO1506: Podpora udržitelnosti centra NTIS - Nové technologie pro informační společnost</a><br>

Návaznosti

P - Projekt vyzkumu a vyvoje financovany z verejnych zdroju (s odkazem do CEP)<br>S - Specificky vyzkum na vysokych skolach<br>I - Institucionalni podpora na dlouhodoby koncepcni rozvoj vyzkumne organizace

Rok uplatnění

2020

Kód důvěrnosti údajů

S - Úplné a pravdivé údaje o projektu nepodléhají ochraně podle zvláštních právních předpisů

Interní identifikační kód produktu

ZCU/KKY/2020/014

Technické parametry

Tento softwarový nástroj používá pokročilé metody strojového učení. Pro bližší informace k technickým detailům a získání licence prosím kontaktujte: Lukáš Picek, KKY FAV Západočeská univerzita v Plzni, Technická 8, 301 00 Plzeň (IČ: 49777513), email: picekl@kky.zcu.cz, tel.: 37763 2125

Ekonomické parametry

Tento SW nástroj je volně dostupný pro využití nekomerčního charakteru. Vzhledem k významnosti výstupu v oblasti vzdělávání a souvisejícímu společenskému dopadu nelze zhodnotit ekonomické parametry SW výstupu - operativní výzkum ve veřejném zájmu.

IČO vlastníka výsledku

49777513

Název vlastníka

Západočeská univerzita v Plzni