BlackBird

This project provides the source code and data for the manuscript "High throughput saliency-based quantification of grape powdery mildew at the microscopic level for disease resistance breeding" published in Oxford Academic/Horticulture Research.

Requirements

• Python 3.6
• PyTorch >= 1.4
• TorchVision
• Captum

Structure

The basic filesystem structure is like what the following tree displays. Essentially, we have two folders code and data to store scripts and dataset, respectively.

Saliency_based_Grape_PM_Quantification/
├── code
│   ├── analysis
│   │   └── workflow
│   ├── classification
│   ├── common
│   ├── figures
│   ├── metric
│   ├── sanity_check
│   ├── script
│   ├── segmentation
│   └── visualization
│       ├── captum
│       └── test_images
├── data
│   └── Hyphal_2019
│       └── images
└── results
    ├── leaf_correlation
    │   └── saliency_based
    └── patch_correlation
        └── saliency_based

More detail explaination of each component in code can be found here

Data

The raw data used in this study can be accessed via this link.

Model

The pretrained model of VGG, ResNet, Inception3, and DeepLab can be accessed via this link.