https://github.com/omidghadami95/efficientnetv2b0---plantvillage

Plant disease detection on PlantVillage dataset using EfficientNetV2-B0
https://github.com/omidghadami95/efficientnetv2b0---plantvillage

confusion-matrix disease-detection efficientnet efficientnetv2 efficientnetv2-b0 efficientnetv2b0 f1-score google-colab googlecolab keras plant-disease-detection plantvillage plantvillage-dataset python tensorflow tensorflow2

Last synced: 22 days ago
JSON representation

Plant disease detection on PlantVillage dataset using EfficientNetV2-B0

• Host: GitHub
• URL: https://github.com/omidghadami95/efficientnetv2b0---plantvillage
• Owner: OmidGhadami95
• Created: 2023-09-11T18:26:48.000Z (over 1 year ago)
• Default Branch: main
• Last Pushed: 2024-03-08T18:48:08.000Z (11 months ago)
• Last Synced: 2024-03-08T20:22:50.768Z (11 months ago)
• Topics: confusion-matrix, disease-detection, efficientnet, efficientnetv2, efficientnetv2-b0, efficientnetv2b0, f1-score, google-colab, googlecolab, keras, plant-disease-detection, plantvillage, plantvillage-dataset, python, tensorflow, tensorflow2
• Language: Jupyter Notebook
• Homepage:
• Size: 1.73 MB
• Stars: 3
• Watchers: 1
• Forks: 1
• Open Issues: 0
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

README

        



# EfficientNetV2B0-PlantVillage

Plant disease detection on PlantVillage dataset using EfficientNetV2-B0.  

In this code, I focused on only 7 labels of this dataset. First PlantVillage dataset has been downloaded and loaded. Then these 7 labels has been seperated for preprocessing step. Processing steps are resizing and augmentation. After preprocessing, I used EfficientNetV2B0 with pre-traind weights of Imagenet dataset. After fine-tuning, for a better analysis, Macro, Micro, and Weighted has been calculated for Precision, Recall, and F1-score. Moreover, the confusion matrix has been shown. At last, EfficientNetV2B0 could achieve around 98% accuracy on these 7 classes.

Note that this dataset is imbalanced, and for better performance in the real world, we need to solve this problem. You can utilize the method that I used in this link for balancing ckplus dataset: 

 https://github.com/OmidGhadami95/EfficientNetV2_Quantization_CK/blob/main/EfficientNetV2_B2_Quantization_CK%2B.ipynb