Create app.py
Browse files
app.py
ADDED
@@ -0,0 +1,47 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
import streamlit as st
|
2 |
+
from ultralyticsplus import YOLO, render_result
|
3 |
+
import PIL.Image as Image
|
4 |
+
import numpy as np
|
5 |
+
import requests
|
6 |
+
from io import BytesIO
|
7 |
+
|
8 |
+
# Initialize Streamlit app
|
9 |
+
st.title("Blood Cell Detection with YOLOv8")
|
10 |
+
|
11 |
+
# Load YOLO model
|
12 |
+
model = YOLO('keremberke/yolov8m-blood-cell-detection')
|
13 |
+
|
14 |
+
# Set model parameters
|
15 |
+
model.overrides['conf'] = 0.25 # NMS confidence threshold
|
16 |
+
model.overrides['iou'] = 0.45 # NMS IoU threshold
|
17 |
+
model.overrides['agnostic_nms'] = False # NMS class-agnostic
|
18 |
+
model.overrides['max_det'] = 1000 # Maximum number of detections per image
|
19 |
+
|
20 |
+
# File uploader for image input
|
21 |
+
uploaded_file = st.file_uploader("Upload an image for detection", type=["jpg", "png"])
|
22 |
+
|
23 |
+
if uploaded_file:
|
24 |
+
# Open the uploaded image
|
25 |
+
image = Image.open(uploaded_file)
|
26 |
+
|
27 |
+
# Perform inference
|
28 |
+
results = model.predict(np.array(image))
|
29 |
+
|
30 |
+
# Display results
|
31 |
+
st.image(image, caption="Uploaded Image", use_column_width=True)
|
32 |
+
|
33 |
+
# Render detection results
|
34 |
+
rendered_image = render_result(model=model, image=image, result=results[0])
|
35 |
+
|
36 |
+
# Show the rendered result
|
37 |
+
st.image(rendered_image, caption="Detection Results", use_column_width=True)
|
38 |
+
|
39 |
+
# Display details of detected boxes
|
40 |
+
st.write("Detection Results:")
|
41 |
+
for box in results[0].boxes:
|
42 |
+
st.write(f"Bounding box: {box.xyxy}")
|
43 |
+
st.write(f"Confidence: {box.conf}")
|
44 |
+
st.write(f"Class: {box.cls}")
|
45 |
+
|
46 |
+
else:
|
47 |
+
st.write("Upload an image to start detection")
|