Update app.py

app.py

@@ -1,106 +1,55 @@
 import gradio as gr
-from ultralytics import YOLO
-import cv2
-import easyocr
-import numpy as np
 from PIL import Image
+import matplotlib.pyplot as plt
+import numpy as np
+from ultralytics import YOLO
 from transformers import pipeline
 
-# Load YOLO model for license plate detection
-model = YOLO('yolo11n-custom.pt')
-model.fuse()
-
-# Load EasyOCR
-reader = easyocr.Reader(['en'])
-
-# Load Transformers pipeline for car damage classification
+# Load models
+license_model = YOLO('yolov8n.pt')  # public YOLOv8 model
 damage_pipe = pipeline("image-classification", model="beingamit99/car_damage_detection")
 
-# --- License Plate Detection Function ---
-def detect_license_plate(image):
-    results = model.predict(image, conf=0.15, iou=0.3, classes=[0])
-    plate_texts = []
-    img_array = np.array(image)
-    img = cv2.cvtColor(img_array, cv2.COLOR_RGB2BGR)
-    img = cv2.cvtColor(img, cv2.COLOR_BGR2RGB)
-
-    img_height, img_width, _ = img.shape
-
-    for result in results:
-        for bbox in result.boxes.xyxy:
-            x1, y1, x2, y2 = map(int, bbox.tolist())
-            plate = img[int(y1):int(y2), int(x1):int(x2)]
-            scale = 2
-            height, width = plate.shape[:2]
-            plate = cv2.resize(plate, (width * scale, height * scale), interpolation=cv2.INTER_CUBIC)
-            lab = cv2.cvtColor(plate, cv2.COLOR_RGB2LAB)
-            l, a, b = cv2.split(lab)
-            clahe = cv2.createCLAHE(clipLimit=3.0, tileGridSize=(8, 8))
-            l = clahe.apply(l)
-            plate = cv2.merge((l, a, b))
-            plate = cv2.cvtColor(plate, cv2.COLOR_LAB2RGB)
-
-            text = reader.readtext(plate, detail=0, allowlist="ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789-")
-            text = " ".join(text).upper()
+# License plate detection
+def detect_license_plate(img: Image.Image):
+    results = license_model.predict(img)
+    result = results[0]
 
-            text_scale = max(1, width / 250)
-            thickness = max(2, width // 200)
-            cv2.rectangle(img, (x1, y1), (x2, y2), (0, 255, 0), thickness)
-            (text_width, text_height), _ = cv2.getTextSize(text, cv2.FONT_HERSHEY_SIMPLEX, text_scale, thickness)
-            text_x = x1 + (width - text_width) // 2
-            text_y = y1 - 10 if y1 > 50 else y2 + text_height + 20
-            text_box_y1 = text_y - text_height - 5
-            text_box_y2 = text_y + 5
-            cv2.rectangle(img, (text_x - 8, text_box_y1 - 3), (text_x + text_width + 8, text_box_y2 + 3), (0, 0, 0), -1)
-            cv2.rectangle(img, (text_x - 5, text_box_y1), (text_x + text_width + 5, text_box_y2), (255, 255, 255), -1)
-            cv2.putText(img, text, (text_x, text_y), cv2.FONT_HERSHEY_SIMPLEX, text_scale, (0, 0, 0), thickness)
+    # Plot boxes
+    img = np.array(img.convert("RGB"))
+    for box in result.boxes.xyxy:
+        x1, y1, x2, y2 = box.int().tolist()
+        img = img.copy()
+        img = cv2.rectangle(img, (x1, y1), (x2, y2), (255, 0, 0), 2)
 
-            plate_texts.append(text)
+    return Image.fromarray(img)
 
-    result_img = Image.fromarray(img)
-    return result_img, "\n".join(plate_texts) if plate_texts else "No license plates detected."
+# Damage classification with bar chart
+def classify_damage(img: Image.Image):
+    predictions = damage_pipe(img)
 
-# --- Car Damage Classification Function ---
-import plotly.graph_objects as go
+    # Prepare bar graph
+    labels = [pred["label"] for pred in predictions]
+    scores = [pred["score"] for pred in predictions]
 
-def classify_car_damage(image):
-    results = damage_pipe(image)
-    labels = [item["label"] for item in results]
-    scores = [item["score"] for item in results]
+    fig, ax = plt.subplots()
+    ax.barh(labels, scores, color='orange')
+    ax.set_xlim(0, 1)
+    ax.set_xlabel("Confidence")
+    ax.set_title("Damage Classification")
+    plt.tight_layout()
 
-    fig = go.Figure(
-        data=[go.Bar(x=labels, y=scores, marker_color="crimson")],
-        layout=go.Layout(
-            title="Car Damage Classification",
-            yaxis=dict(title="Confidence Score", range=[0, 1]),
-            xaxis=dict(title="Damage Type"),
-            template="plotly_white"
-        )
-    )
+    return fig
 
-    return image, fig
+# Combined interface
+with gr.Blocks() as app:
+    gr.Markdown("# 🚗 Car License Plate & Damage Detector")
 
-# --- Gradio Interface ---
-plate_tab = gr.Interface(
-    fn=detect_license_plate,
-    inputs=gr.Image(type="pil"),
-    outputs=[
-        gr.Image(type="pil", label="Detected Image"),
-        gr.Textbox(label="Detected License Plates")
-    ],
-    title="🚘 License Plate Detector",
-    description="Upload an image with license plates to detect them using YOLO and EasyOCR."
-)
+    with gr.Row():
+        image_input = gr.Image(type="pil", label="Upload Car Image")
+        license_output = gr.Image(label="License Plate Detection")
+        damage_output = gr.Plot(label="Damage Classification")
 
-damage_tab = gr.Interface(
-    fn=classify_car_damage,
-    inputs=gr.Image(type="pil", label="Upload Damaged Car Image"),
-    outputs=[
-        gr.Image(type="pil", label="Car Image"),
-        gr.Plot(label="Damage Classification")
-    ],
-    title="💥 Car Damage Detector",
-    description="Classifies car damage severity (minor, moderate, severe)."
-)
+    image_input.change(fn=detect_license_plate, inputs=image_input, outputs=license_output)
+    image_input.change(fn=classify_damage, inputs=image_input, outputs=damage_output)
 
-gr.TabbedInterface([plate_tab, damage_tab], ["License Plate Detector", "Car Damage Detector"]).launch()
+app.launch()