Update app.py

app.py

@@ -1,148 +1,60 @@
-import numpy as np
-import cv2
-import matplotlib.pyplot as plt
 import gradio as gr
-import tempfile
-import os
-import json
-
-def calculate_snr(image, roi_json):
-    roi_points = json.loads(roi_json)
-
-    # 创建临时文件来保存上传的图像
-    with tempfile.NamedTemporaryFile(delete=False, suffix='.png') as temp_file:
-        temp_filename = temp_file.name
-        cv2.imwrite(temp_filename, image)
-
-    # 读取图像
-    img = cv2.imread(temp_filename, cv2.IMREAD_GRAYSCALE)
-
-    # 计算信号和背景区域
-    signal_mask = np.zeros(img.shape, np.uint8)
-    cv2.fillPoly(signal_mask, [np.array(roi_points)], 255)
-    signal = cv2.bitwise_and(img, img, mask=signal_mask)
-
-    background_mask = cv2.bitwise_not(signal_mask)
-    background = cv2.bitwise_and(img, img, mask=background_mask)
-
-    # 计算信号区域的平均强度
-    signal_mean = np.mean(signal[signal_mask == 255])
-
-    # 计算背景区域的标准差
-    background_std = np.std(background[background_mask == 255])
-
-    # 计算SNR
-    snr = signal_mean / background_std if background_std != 0 else float('inf')
+import numpy as np
+from PIL import Image, ImageDraw
+import io
 
-    # 创建图像结果
-    plt.figure(figsize=(10, 5))
-    plt.subplot(121)
-    plt.imshow(signal, cmap='gray')
-    plt.title('Signal ROI')
-    plt.subplot(122)
-    plt.imshow(background, cmap='gray')
-    plt.title('Background ROI')
+def calculate_snr(image, roi_mask):
+    # Convert image to numpy array
+    img_array = np.array(image)
     
-    # 保存图像结果到临时文件
-    result_filename = tempfile.mktemp(suffix='.png')
-    plt.savefig(result_filename)
-    plt.close()
-
-    # 删除临时文件
-    os.unlink(temp_filename)
-
-    return (
-        f"信号平均强度: {signal_mean:.2f}\n"
-        f"背景标准差: {background_std:.2f}\n"
-        f"信噪比(SNR): {snr:.2f}",
-        result_filename
-    )
-
-# 自定义HTML组件和JavaScript代码
-html_code = """
-<div>
-    <canvas id="roi-canvas" width="500" height="500" style="border:1px solid #000000;"></canvas>
-    <button onclick="startDrawing()">Start Drawing</button>
-    <button onclick="stopDrawing()">Stop Drawing</button>
-    <button onclick="clearCanvas()">Clear</button>
-    <textarea id="roi-points" style="display:none;"></textarea>
-</div>
-<script>
-    let canvas = document.getElementById('roi-canvas');
-    let ctx = canvas.getContext('2d');
-    let roi = [];
-    let drawing = false;
-
-    canvas.addEventListener('mousedown', (e) => {
-        drawing = true;
-        roi.push({x: e.offsetX, y: e.offsetY});
-    });
-
-    canvas.addEventListener('mousemove', (e) => {
-        if (drawing) {
-            let lastPoint = roi[roi.length - 1];
-            ctx.beginPath();
-            ctx.moveTo(lastPoint.x, lastPoint.y);
-            ctx.lineTo(e.offsetX, e.offsetY);
-            ctx.strokeStyle = 'green';
-            ctx.lineWidth = 2;
-            ctx.stroke();
-            ctx.closePath();
-            roi.push({x: e.offsetX, y: e.offsetY});
-        }
-    });
-
-    canvas.addEventListener('mouseup', () => {
-        drawing = false;
-        document.getElementById('roi-points').value = JSON.stringify(roi);
-    });
-
-    function startDrawing() {
-        drawing = true;
-    }
-
-    function stopDrawing() {
-        drawing = false;
-    }
-
-    function clearCanvas() {
-        ctx.clearRect(0, 0, canvas.width, canvas.height);
-        roi = [];
-        document.getElementById('roi-points').value = '';
-    }
-</script>
-"""
-
-# 创建Gradio接口
-with gr.Blocks() as demo:
-    with gr.Row():
-        image_input = gr.Image(label="上传图像")
-        roi_output = gr.HTML(html_code, label="绘制ROI区域")
-        roi_data = gr.Textbox(label="ROI Data", visible=False)
-
-    calculate_button = gr.Button("计算SNR")
-    result_text = gr.Textbox(label="结果")
-    result_image = gr.Image(label="ROI 图像")
-
-    def extract_roi_data(html_output):
-        return html_output
-
-    image_input.change(
-        fn=lambda x: "",
-        inputs=image_input,
-        outputs=roi_output
-    )
-
-    roi_output.change(
-        fn=extract_roi_data,
-        inputs=roi_output,
-        outputs=roi_data
-    )
-
-    calculate_button.click(
-        fn=calculate_snr,
-        inputs=[image_input, roi_data],
-        outputs=[result_text, result_image]
-    )
+    # Create ROI and background masks
+    roi = img_array * roi_mask
+    bg = img_array * (1 - roi_mask)
+    
+    # Calculate mean and standard deviation
+    roi_mean = np.mean(roi[roi != 0])
+    bg_mean = np.mean(bg[bg != 0])
+    bg_std = np.std(bg[bg != 0])
+    
+    # Calculate SNR
+    snr = (roi_mean - bg_mean) / bg_std
+    
+    return snr
 
-demo.launch()
+def process_image(input_image, roi_coordinates):
+    image = Image.open(io.BytesIO(input_image))
+    
+    # Create a mask for the ROI
+    mask = Image.new('L', image.size, 0)
+    draw = ImageDraw.Draw(mask)
+    draw.polygon(roi_coordinates, fill=255)
+    
+    # Convert mask to binary numpy array
+    mask_array = np.array(mask) / 255
+    
+    # Calculate SNR
+    snr = calculate_snr(image.convert('L'), mask_array)
+    
+    # Visualize the ROI on the image
+    overlay = Image.new('RGBA', image.size, (0, 0, 0, 0))
+    draw = ImageDraw.Draw(overlay)
+    draw.polygon(roi_coordinates, fill=(255, 0, 0, 64))
+    result_image = Image.alpha_composite(image.convert('RGBA'), overlay)
+    
+    return result_image, f"SNR: {snr:.2f}"
+
+iface = gr.Interface(
+    fn=process_image,
+    inputs=[
+        gr.Image(type="numpy", label="Upload Fluorescence Image"),
+        gr.ImageMask(label="Draw ROI", source="upload")
+    ],
+    outputs=[
+        gr.Image(type="pil", label="Result"),
+        gr.Textbox(label="Signal-to-Noise Ratio (SNR)")
+    ],
+    title="Fluorescence Staining Quality Assessment",
+    description="Upload a fluorescence image and draw the region of interest (ROI) to calculate the Signal-to-Noise Ratio (SNR)."
+)
+
+iface.launch()