BluVisionMicro

We have developed the BluVision image analysis Framework for studying plant-pathogen interactions on microscopic level. The system is build to study the life cycle of the important barley and wheat pathogen powdery mildew by collecting and analyzing image data from three key developmental stages.

BluVision is a software and hardware framework for high-throughput image acquisition and analysis of microscopic images in plant pathology. The microscopic module is based on a state of the art high-throughput microscope scanner (Zeiss Axio Scan.Z1).

Figure 1: Zeiss Axio Scan.Z1

Our image analysis pipeline is aimed to detect and analyze microscopic infection events (Figure 2 & 3).

Figure 2: BluVision Micro Module

Figure 3: Blumeria graminis hyphae detetcion

Installation

Tip

We recommend to install Anaconda and for managing dependencies, it is often recommended to create a new environment for your project:

Install Anaconda from https://www.anaconda.com/distribution/

Open the Anaconda Prompt

conda create --name bluvisionmicro_env python=3.8
conda activate bluvisionmicro_env

Important

Clone the Github repository:

git clone https://github.com/snowformatics/BluVisionMicro.git

Important

Install dependencies:

pip install requiremenmts.txt

Important

Download and copy the CNN model:

https://github.com/snowformatics/BluVisionMicro/releases/download/v1.0.0/09112020_1.h5

Run the analysis

Important

Move to the BluVisionMicro folder and make sure your environment is activated

For small colonies (< 50 hai) type the command:
python cli.py -s source_path -d destination_path -p mildew_small -m analysis -se 0.05

For large colonies (> 50 hai) type the command:
python cli.py -s source_path -d destination_path -p mildew_large -m analysis -se 0.05

Export the results

python cli.py -s source_path -d destination_path -p mildew -m results

Parameters

Parameter	Description
-s	Path to source CZI images.
-d	Path to store the ROIs.
-p	Pathogen.
-m	Mode, analysis or results.
-se	Sensitivity for the CNN to predict hyphae. We recommend using strict values for host interactions (se = 0.0) and relaxed values for nonhost interactions (0.05).