Upload 6 files

README.md

@@ -1,12 +1,13 @@
----
-title: NajmPOC
-emoji: 👁
-colorFrom: blue
-colorTo: pink
-sdk: gradio
-sdk_version: 5.13.1
-app_file: app.py
-pinned: false
----
-
-Check out the configuration reference at https://huggingface.co/docs/hub/spaces-config-reference
+---
+title: IDs Demo
+emoji: 🐢
+colorFrom: yellow
+colorTo: purple
+sdk: gradio
+sdk_version: 5.12.0
+app_file: app.py
+pinned: false
+short_description: This is a demo for information extraction from Egyption IDs
+---
+
+Check out the configuration reference at https://huggingface.co/docs/hub/spaces-config-reference

app.py

@@ -0,0 +1,102 @@
+import gradio as gr
+from recognization.recognization import TextRecognition
+from detection.recognize_id.detect_and_recognize_id import Recognize_ID
+from detection.detection import detection
+import os
+import numpy as np
+import cv2
+
+# Define a function to crop the image based on bounding box coordinates
+def crop_image(image, x, y, w, h):
+    """
+    Crop the image based on the provided bounding box coordinates.
+    x, y are the top-left coordinates, w is width, and h is height.
+    """
+    # Convert to integer values
+    x, y, w, h = int(x), int(y), int(w), int(h)
+    
+    # Convert the image to a numpy array if it's in PIL format (e.g., if it came from Gradio)
+    if isinstance(image, np.ndarray):
+        img = image
+    else:
+        img = np.array(image)
+
+    # Crop the image using the coordinates
+    cropped_img = img[y:y+h, x:x+w]
+    return cropped_img
+
+# Define a dummy prediction function for license card data extraction
+def predict_image(image):
+
+    # Recognize ID (adjust for license cards as needed)
+    rec_id = Recognize_ID()
+    id = rec_id.give_me_id_number(image)
+
+    # Detection (update the detection method to work with license card layout)
+    det = detection()
+    detection_list = det.full_pipeline(image)
+
+    result = ''
+    # Loop through the detection list, extracting text from the new fields (update to use the license card coordinates)
+    recognizer = TextRecognition()
+
+    # Bounding box classes based on provided coordinates
+    bounding_boxes = [
+        (0.693423, 0.231959, 0.249721, 0.055670),
+        (0.692308, 0.288660, 0.251951, 0.045361),
+        (0.731327, 0.350515, 0.158305, 0.057732),
+        (0.310479, 0.354639, 0.166109, 0.041237),
+        (0.608696, 0.426804, 0.405797, 0.070103),
+        (0.749721, 0.502062, 0.132664, 0.059794),
+        (0.737458, 0.558763, 0.139353, 0.053608),
+        (0.296544, 0.554639, 0.066890, 0.057732),
+        (0.672798, 0.808247, 0.143813, 0.049485)
+    ]
+
+    # Convert the image to numpy array if necessary (e.g., if it's a PIL image)
+    if isinstance(image, np.ndarray):
+        img = image
+    else:
+        img = np.array(image)
+
+    # Extract and recognize text from the specific bounding boxes
+    for bbox in bounding_boxes:
+        x, y, w, h = bbox
+        # Convert relative coordinates to absolute pixel values (assuming image dimensions are available)
+        h, w_img = img.shape[:2]
+        x_abs = int(x * w_img)
+        y_abs = int(y * h)
+        w_abs = int(w * w_img)
+        h_abs = int(h * h)
+
+        # Crop the image using OpenCV
+        cropped_image = crop_image(img, x_abs, y_abs, w_abs, h_abs)
+        
+        # Recognize text in the cropped image
+        recognized_word = recognizer.recognize_image(cropped_image)
+        result = result + recognized_word + ' '
+        result += '\n'
+
+    # Add ID number
+    result = result + id
+
+    return result
+
+# List of paths to your sample images
+current_dir = os.path.dirname(os.path.abspath(__file__))
+sample_images = [
+    os.path.join(current_dir, "samples/license_card_sample.png")  # Update to your sample license card image
+]
+
+# Create the Gradio interface
+interface = gr.Interface(
+    fn=predict_image,  # Function to run    
+    inputs="image",    # Input type
+    outputs="text",    # Output type
+    title="Information extraction from License Cards",  # Updated title for license cards
+    description="Upload a license card image to extract data.",
+    examples=sample_images
+)
+
+# Launch the app
+interface.launch()

detection.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:99ec8e866ec9193070894776a3a9183b1fc51fcc5b374abb26f72585bf31266a
+size 40513509

recognization_id.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:929c9b568d2c4ece0a993d5f5d327ccac21a0edc1d91bdee904d27a590e87dcd
+size 57156628

recognization_model.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3a12da308d75a0b39d0f8b6691b8cd4a0956375d88e0e8fe0443ff9855f45737
+size 189602122

requirements.txt

@@ -0,0 +1,6 @@
+torch
+numpy
+tqdm
+opencv-python
+ultralytics
+