PDR based on modified PlantVillage Dataset
About the dataset: This is modified version of PlantVillage Dataset. This dataset has been modified to work with data generators of Keras library. The annotations has been modified so that labelling is easier. The train-test split is set at 85:15. Link for modified dataset: https://drive.google.com/file/d/1Mj6wsKBZN2ycAyyIMs2lI361deuCJqBI/view?usp=sharing
Description: There are 5 ipython notebooks that are aimed at recognizing plant diseases. There is one notebook each for VGG19, Xception, InceptionV3, ResNet105V2, and EfficientNetB7. Additionally, to the pretrained models, the following layers have been added: 1 Batch normalization layer, 1 Dense layers of 512 nodes, and 1 Dense layer for softmax classification.
Keys for alternate annotations:
key= {'00': 'Apple___Apple_scab', '01': 'Apple___Black_rot', '02': 'Apple___Cedar_apple_rust', '03': 'Apple___healthy', '04': 'Blueberry___healthy', '05': 'Cherry_(including_sour)__Powdery_mildew', '06': 'Cherry(including_sour)__healthy', '07': 'Corn(maize)__Cercospora_leaf_spot Gray_leaf_spot', '08': 'Corn(maize)_Common_rust', '09': 'Corn(maize)__Northern_Leaf_Blight', '10': 'Corn(maize)healthy', '11': 'Grape___Black_rot', '12': 'Grape___Esca(Black_Measles)', '13': 'Grape___Leaf_blight(Isariopsis_Leaf_Spot)', '14': 'Grape___healthy', '15': 'Orange___Haunglongbing(Citrus_greening)', '16': 'Peach___Bacterial_spot', '17': 'Peach___healthy', '18': 'Pepper,_bell___Bacterial_spot', '19': 'Pepper,_bell___healthy', '20': 'Potato___Early_blight', '21': 'Potato___Late_blight', '22': 'Potato___healthy', '23': 'Raspberry___healthy', '24': 'Soybean___healthy', '25': 'Squash___Powdery_mildew', '26': 'Strawberry___Leaf_scorch', '27': 'Strawberry___healthy', '28': 'Tomato___Bacterial_spot', '29': 'Tomato___Early_blight', '30': 'Tomato___Late_blight', '31': 'Tomato___Leaf_Mold', '32': 'Tomato___Septoria_leaf_spot', '33': 'Tomato___Spider_mites Two-spotted_spider_mite', '34': 'Tomato___Target_Spot', '35': 'Tomato___Tomato_Yellow_Leaf_Curl_Virus', '36': 'Tomato___Tomato_mosaic_virus', '37': 'Tomato___healthy'}
No. of sample, belonging to total 38 classes. 85:15 split for Train:Test.
Train: 46141 images && Test: 8162 images
(Time using colab GPU)
| DNN | No. of layers | Code Available | Average Time per epoch |
|---|---|---|---|
| 1. VGG19 | 26 | Yes | 766 seconds OR 12 min 46 sec |
| 2. Xception | 136 | Yes | |
| 3. InceptionV3 | 315 | Yes | |
| 4. ResNet152V2 | 568 | Yes | 1167 seconds OR 19 min 27 sec |
| 5. EfficientNetB7 | 817 | Yes |
Target Epochs for each DNN: 100 Epoch
H5 files will be added later.
(Futher Details will be added.)