https://github.com/pitmonticone/plantdiseaseclassification

Dataset Analysis & CNN Models Optimization for Plant Disease Classification.
https://github.com/pitmonticone/plantdiseaseclassification

classification-algorithims computer-science computer-vision convolutional-neural-networks data-science deep-learning deep-neural-networks neural-network neural-networks plant-disease-classification plant-disease-detection plant-pathology-fgvc7

Last synced: 11 days ago
JSON representation

Dataset Analysis & CNN Models Optimization for Plant Disease Classification.

• Host: GitHub
• URL: https://github.com/pitmonticone/plantdiseaseclassification
• Owner: pitmonticone
• License: mit
• Created: 2020-08-29T22:46:22.000Z (over 4 years ago)
• Default Branch: master
• Last Pushed: 2021-11-12T14:08:41.000Z (about 3 years ago)
• Last Synced: 2024-11-21T06:38:33.240Z (2 months ago)
• Topics: classification-algorithims, computer-science, computer-vision, convolutional-neural-networks, data-science, deep-learning, deep-neural-networks, neural-network, neural-networks, plant-disease-classification, plant-disease-detection, plant-pathology-fgvc7
• Language: HTML
• Homepage: https://pitmonticone.github.io/PlantDiseaseClassification/
• Size: 22.3 MB
• Stars: 14
• Watchers: 3
• Forks: 5
• Open Issues: 0
• Metadata Files:
  • Readme: README.md
  • License: LICENSE

Awesome Lists containing this project

README

        




    

  

  

  

  

  

  

  

  

    

  

  

  

  

  





  Plant Disease Classification





  Dataset Analysis and CNN Models Optimization 





  

    

  

  

    

  

  

    

  

  



## How to Explore this Work

* Read the report in [PDF](https://pitmonticone.github.io/PlantDiseaseClassification/Report/report.pdf) or [HTML](https://pitmonticone.github.io/PlantDiseaseClassification/Report/report.html) format.

* Read the code in the [Jupyter notebook](https://nbviewer.jupyter.org/github/pitmonticone/PlantDiseaseClassification/blob/master/Notebooks/notebook.ipynb).

* Run the code in the [Kaggle notebook](https://www.kaggle.com/inphyt2020/neuralnetworksproject).

## Extended Abstract

### Problem 

Misdiagnosis of the many diseases impacting agricultural crops can lead to misuse of chemicals leading to the emergence of resistant pathogen strains, increased input costs, and more outbreaks with significant economic loss and environmental impacts. Current disease diagnosis based on human scouting is time-consuming and expensive, and although computer-vision based models have the promise to increase efficiency, the great variance in symptoms due to age of infected tissues, genetic variations, and light conditions within trees decreases the accuracy of detection.

### Objectives

The [Plant Pathology Challenge](https://www.kaggle.com/c/plant-pathology-2020-fgvc7/overview) we have attended consists in training a model using images of the [training dataset](https://arxiv.org/abs/2004.11958) to

* accurately classify a given image from testing dataset into different diseased category or a healthy leaf; 

* accurately distinguish between many diseases, sometimes more than one on a single leaf;

* deal with rare classes and novel symptoms;

* address depth perception—angle, light, shade, physiological age of the leaf; 

* Incorporate expert knowledge in identification, annotation, quantification, and guiding computer vision to search for relevant features during learning.

Submissions are evaluated on [mean column-wise ROC AUC](https://www.kaggle.com/c/plant-pathology-2020-fgvc7/overview/evaluation).

### Data 

Both the training and the testing [datasets](https://www.kaggle.com/c/plant-pathology-2020-fgvc7/data) are composed of 1821 high-quality, real-life symptom images of multiple apple foliar diseases to be classified into four categories: `healthy`, `multiple_diseases`, `rust`, `scab`. 

### Methods

1. Class balancing with [`SMOTE`](https://imbalanced-learn.readthedocs.io/en/stable/generated/imblearn.over_sampling.SMOTE.html)

1. Data augmentation with [Keras `ImageDataGenerator`](https://keras.io/api/preprocessing/image/)

1. Learning rate schedule

1. Optimal dropout

1. Epoch grid search

1. Visualization of convolutional filters and activation maps of the layers.

### Results 

* **ROC = 0.972** applying the pre-trained Keras model [`DenseNet121`](https://keras.io/api/applications/densenet/#densenet121-function).

* **ROC = 0.937** applying a CNN which has been defined and trained from scratch. 

### References 

1. [Plant Pathology 2020 - FGVC7: Identify the category of foliar diseases in apple trees](https://www.kaggle.com/c/plant-pathology-2020-fgvc7), *Kaggle* (2020). 

1. Ranjita Thapa et al. [The Plant Pathology 2020 challenge dataset to classify foliar disease of apples](https://arxiv.org/abs/2004.11958), *arXiv pre-print* (2020).