FungAI -

Automatic Fungal Identification Based On Deep Learning

The FungAI-V1.1 server provides an automated prediction system for fungal species identification.
This AI-based model leverages deep learning techniques applied to time-lapse images.

The server identifies the top five closest classifications to the related fungal species.

The current version was trained on 26,451 high-resolution images representing 110 species from 35 genera derived from fungal strains in the IBT Culture Collection.


Fungi Image1

Overview of genus

This section gives an overview of the genus in focus. It includes details about classification and characteristics.

Fungi Image2

Overview of species

This section provides an overview of different species within the genus, highlighting their unique features and relevance.

Instructions


1) The input images must be in 6-well plates, as shown below:

IBT 23255

IBT 23255

2) Choose your file and click Upload. You will then see the five closest classifications to your input. (Below is a short video demonstrating how the model is used).

FungAI


Upload an image to classify:

No file chosen

Sample Data


IBT 23255

IBT 26955

IBT 32802

IBT 41508

Team


Person 1
Marjan Mansourvar marjma@dtu.dk
Computer Vision, AI Model Development, and Deep Learning Approaches
Person 2
Jakob Blæsbjerg Hoof jblni@dtu.dk
IBT Culture Collection
Person 3
Steen Smidth Brewer s182833@student.dtu.dk
Research Assistant
Person 4
Rasmus Johansen Rieneck s183734@student.dtu.dk
AI-student, Web Application Development, Frontend Frameworks and Backend systems