Create app.py

app.py

@@ -0,0 +1,89 @@
+import gradio as gr
+from PIL import Image, ImageDraw, ImageFont
+import os
+from io import BytesIO
+from roboflow import Roboflow
+
+rf = Roboflow(api_key=roboflow_key)
+project = rf.workspace('yudi-pratama-putra-rwuep').project("corn-pest-detection-2")
+model = project.version(3).model
+
+
+def predict_image(image, confidence, overlap):
+    colors = [
+        "Red",
+        "Green",
+        "Blue",
+        "Yellow",
+        "Cyan",
+        "Magenta",
+        "Orange",
+        "Purple",
+        "Brown",
+        "Pink",
+        "DarkRed",
+        "Black",
+        "White"
+    ]
+
+    pest_class = ['aphids', 'army worm', 'black cutworm', 'corn borer', 'grub', 'large cutworm', 'mole cricket', 'peach borer', 'potosiabre vitarsis', 'red spider', 'white margined moth', 'wireworm', 'yellow cutworm']
+
+    prediction = model.predict(image, confidence=confidence, overlap=overlap).json()
+
+    img = Image.open(image).convert("RGB")
+
+    img = img.resize((img.width * 2, img.height * 2))
+
+    draw = ImageDraw.Draw(img)
+    font = ImageFont.load_default(size=30)
+
+    for result in prediction['predictions']:
+        x0 = result['x'] - result['width'] / 2
+        y0 = result['y'] - result['height'] / 2
+        x1 = result['x'] + result['width'] / 2
+        y1 = result['y'] + result['height'] / 2
+
+        x0 *= 2
+        y0 *= 2
+        x1 *= 2
+        y1 *= 2
+
+        for i in range(len(pest_class)):
+          if result['class'] == pest_class[i]:
+              draw.rectangle([x0, y0, x1, y1], outline=colors[i], width=3)
+
+              label = f"{result['class']} ({result['confidence']*100:.2f}%)"
+
+              draw.text((x0, y0 - 35), label, fill=colors[i], font=font)
+
+    if prediction['predictions'] == []:
+        img = Image.open(image)
+
+    return img
+
+inputs_image = [
+    gr.Image(type='filepath', label='input image'),
+    gr.Slider(minimum=0, maximum=100, value=40, label='Confidence (%)'),
+    gr.Slider(minimum=0, maximum=100, value=30, label='Overlap (%)')
+]
+outputs_image = [
+    gr.Image(type='numpy', label='output image')
+]
+
+interface_image = gr.Interface(
+    fn=predict_image,
+    inputs=inputs_image,
+    outputs=outputs_image,
+    title="Corn Pest Detection",
+    description=(
+        "Upload an image and the model will detect pests.\n\n"
+        "Model detected: aphids, army worm, black cutworm, corn borer, grub, large cutworm, mole cricket, peach borer, "
+        "potosiabre vitarsis, red spider, white margined moth, wireworm, yellow cutworm.\n\n"
+        "Confidence: The higher the confidence value, the more certain the model is about the detected object being correct. "
+        "For example, a higher confidence threshold will filter out less certain predictions.\n\n"
+        "Overlap: The higher the accepted overlap value, the more predictions are allowed, even if they overlap with each other. "
+        "A higher overlap value can help detect multiple objects that are close together."
+    )
+)
+
+interface_image.launch(share=True)