Spaces:
Runtime error
Runtime error
Update app.py
Browse files
app.py
CHANGED
@@ -0,0 +1,110 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
import gradio as gr
|
2 |
+
import tensorflow as tf
|
3 |
+
import cv2
|
4 |
+
import matplotlib.pyplot as plt
|
5 |
+
import numpy as np
|
6 |
+
|
7 |
+
# Load the model
|
8 |
+
model = tf.keras.models.load_model('plant_model_v5-beta.h5')
|
9 |
+
|
10 |
+
# Define the class names
|
11 |
+
class_names = {
|
12 |
+
0: 'Apple___Apple_scab',
|
13 |
+
1: 'Apple___Black_rot',
|
14 |
+
2: 'Apple___Cedar_apple_rust',
|
15 |
+
3: 'Apple___healthy',
|
16 |
+
4: 'Not a plant',
|
17 |
+
5: 'Blueberry___healthy',
|
18 |
+
6: 'Cherry___Powdery_mildew',
|
19 |
+
7: 'Cherry___healthy',
|
20 |
+
8: 'Corn___Cercospora_leaf_spot Gray_leaf_spot',
|
21 |
+
9: 'Corn___Common_rust',
|
22 |
+
10: 'Corn___Northern_Leaf_Blight',
|
23 |
+
11: 'Corn___healthy',
|
24 |
+
12: 'Grape___Black_rot',
|
25 |
+
13: 'Grape___Esca_(Black_Measles)',
|
26 |
+
14: 'Grape___Leaf_blight_(Isariopsis_Leaf_Spot)',
|
27 |
+
15: 'Grape___healthy',
|
28 |
+
16: 'Orange___Haunglongbing_(Citrus_greening)',
|
29 |
+
17: 'Peach___Bacterial_spot',
|
30 |
+
18: 'Peach___healthy',
|
31 |
+
19: 'Pepper,_bell___Bacterial_spot',
|
32 |
+
20: 'Pepper,_bell___healthy',
|
33 |
+
21: 'Potato___Early_blight',
|
34 |
+
22: 'Potato___Late_blight',
|
35 |
+
23: 'Potato___healthy',
|
36 |
+
24: 'Raspberry___healthy',
|
37 |
+
25: 'Soybean___healthy',
|
38 |
+
26: 'Squash___Powdery_mildew',
|
39 |
+
27: 'Strawberry___Leaf_scorch',
|
40 |
+
28: 'Strawberry___healthy',
|
41 |
+
29: 'Tomato___Bacterial_spot',
|
42 |
+
30: 'Tomato___Early_blight',
|
43 |
+
31: 'Tomato___Late_blight',
|
44 |
+
32: 'Tomato___Leaf_Mold',
|
45 |
+
33: 'Tomato___Septoria_leaf_spot',
|
46 |
+
34: 'Tomato___Spider_mites Two-spotted_spider_mite',
|
47 |
+
35: 'Tomato___Target_Spot',
|
48 |
+
36: 'Tomato___Tomato_Yellow_Leaf_Curl_Virus',
|
49 |
+
37: 'Tomato___Tomato_mosaic_virus',
|
50 |
+
38: 'Tomato___healthy'
|
51 |
+
}
|
52 |
+
|
53 |
+
def edge_and_cut(img, threshold1, threshold2):
|
54 |
+
emb_img = img.copy()
|
55 |
+
edges = cv2.Canny(img, threshold1, threshold2)
|
56 |
+
edge_coors = []
|
57 |
+
for i in range(edges.shape[0]):
|
58 |
+
for j in range(edges.shape[1]):
|
59 |
+
if edges[i][j] != 0:
|
60 |
+
edge_coors.append((i, j))
|
61 |
+
|
62 |
+
if len(edge_coors) == 0:
|
63 |
+
return emb_img
|
64 |
+
|
65 |
+
row_min = edge_coors[np.argsort([coor[0] for coor in edge_coors])[0]][0]
|
66 |
+
row_max = edge_coors[np.argsort([coor[0] for coor in edge_coors])[-1]][0]
|
67 |
+
col_min = edge_coors[np.argsort([coor[1] for coor in edge_coors])[0]][1]
|
68 |
+
col_max = edge_coors[np.argsort([coor[1] for coor in edge_coors])[-1]][1]
|
69 |
+
new_img = img[row_min:row_max, col_min:col_max]
|
70 |
+
|
71 |
+
emb_color = np.array([255], dtype=np.uint8) # Grayscale version of [255, 0, 0]
|
72 |
+
emb_img[row_min-10:row_min+10, col_min:col_max] = emb_color
|
73 |
+
emb_img[row_max-10:row_max+10, col_min:col_max] = emb_color
|
74 |
+
emb_img[row_min:row_max, col_min-10:col_min+10] = emb_color
|
75 |
+
emb_img[row_min:row_max, col_max-10:col_max+10] = emb_color
|
76 |
+
|
77 |
+
return emb_img
|
78 |
+
|
79 |
+
|
80 |
+
|
81 |
+
def classify_and_visualize(image):
|
82 |
+
# Preprocess the image
|
83 |
+
img_array = tf.image.resize(image, [256, 256])
|
84 |
+
img_array = tf.expand_dims(img_array, 0) / 255.0
|
85 |
+
|
86 |
+
# Make a prediction
|
87 |
+
prediction = model.predict(img_array)
|
88 |
+
predicted_class = tf.argmax(prediction[0], axis=-1)
|
89 |
+
confidence = np.max(prediction[0])
|
90 |
+
|
91 |
+
if confidence < 0.60:
|
92 |
+
class_name = "The image you uploaded might not be in the dataset. Try making your leaf background white."
|
93 |
+
bounded_image = image
|
94 |
+
else:
|
95 |
+
class_name = class_names[predicted_class.numpy()]
|
96 |
+
bounded_image = edge_and_cut(image, 200, 400)
|
97 |
+
|
98 |
+
return class_name, confidence, bounded_image
|
99 |
+
|
100 |
+
|
101 |
+
iface = gr.Interface(
|
102 |
+
fn=classify_and_visualize,
|
103 |
+
inputs="image",
|
104 |
+
outputs=["text", "number", "image"],
|
105 |
+
interpretation="default",
|
106 |
+
examples=[
|
107 |
+
['examples/grot.jpg'],
|
108 |
+
['examples/corn.jpg'],
|
109 |
+
])
|
110 |
+
iface.launch()
|