Spaces:

Fortx69
/

face

Runtime error

face / app.py

Create app.py

f239a3f verified almost 2 years ago

4.18 kB

import os
import pandas as pd
from datetime import datetime
from deepface import DeepFace
import gradio as gr
from PIL import Image

# Ensure necessary directories exist
os.makedirs("registered_faces", exist_ok=True)
os.makedirs("attendance", exist_ok=True)

# Adjust the similarity threshold
SIMILARITY_THRESHOLD = 0.4  # Lower means stricter; increase for more flexibility

# 1. Register a face
def register_face(image, name):
    if not name.strip():
        return "Please provide a valid name."
    
    file_path = f"registered_faces/{name}.jpg"
    image.save(file_path)
    return f"Face registered for {name}!"

# 2. Recognize a face and mark attendance
def recognize_face(image):
    if not os.listdir("registered_faces"):
        return "No registered faces found. Please register a face first."

    recognized_name = "Unknown"
    debug_logs = []  # Store debug information

    # Save the uploaded image temporarily
    temp_image_path = "temp_recognition.jpg"
    image.save(temp_image_path)

    for registered_file in os.listdir("registered_faces"):
        registered_path = f"registered_faces/{registered_file}"
        try:
            # Perform face verification
            result = DeepFace.verify(
                img1_path=registered_path, 
                img2_path=temp_image_path, 
                enforce_detection=True,  # Ensure face detection
                model_name="Facenet",  # Use a different model for testing
                distance_metric="cosine"  # Use cosine distance for better comparison
            )
            similarity = result["distance"]
            debug_logs.append(f"Compared with {registered_file}: Verified={result['verified']}, Distance={similarity}")

            # If a match is found
            if similarity < SIMILARITY_THRESHOLD:
                recognized_name = os.path.splitext(registered_file)[0]
                break
        except Exception as e:
            debug_logs.append(f"Error comparing with {registered_file}: {str(e)}")
            continue

    # Remove temporary file after processing
    if os.path.exists(temp_image_path):
        os.remove(temp_image_path)

    # Log attendance if recognized
    if recognized_name != "Unknown":
        now = datetime.now().strftime("%Y-%m-%d %H:%M:%S")
        attendance_file = "attendance/attendance.csv"

        if not os.path.exists(attendance_file):
            df = pd.DataFrame(columns=["Name", "Time"])
        else:
            df = pd.read_csv(attendance_file)
        
        df = pd.concat([df, pd.DataFrame({"Name": [recognized_name], "Time": [now]})], ignore_index=True)
        df.to_csv(attendance_file, index=False)

    # Debug output
    return f"Recognized: {recognized_name}\nDebug Logs:\n" + "\n".join(debug_logs)

# 3. Download attendance records
def download_attendance():
    attendance_file = "attendance/attendance.csv"
    if os.path.exists(attendance_file):
        return attendance_file
    else:
        return "No attendance records found."

# 4. Delete all registered faces
def delete_registered_faces():
    folder = "registered_faces"
    if os.listdir(folder):
        for file in os.listdir(folder):
            os.remove(os.path.join(folder, file))
        return "All registered faces have been deleted."
    return "No registered faces to delete."

# Gradio interfaces
register_interface = gr.Interface(
    fn=register_face,
    inputs=[gr.Image(type="pil"), gr.Text(label="Name")],
    outputs="text",
    title="Register Face"
)

recognize_interface = gr.Interface(
    fn=recognize_face,
    inputs=gr.Image(type="pil"),  # Corrected to use "pil"
    outputs="text",
    title="Recognize Face"
)

download_interface = gr.Interface(
    fn=download_attendance,
    inputs=None,
    outputs="file",
    title="Download Attendance"
)

delete_interface = gr.Interface(
    fn=delete_registered_faces,
    inputs=None,
    outputs="text",
    title="Delete Registered Faces"
)

# Tabbed Gradio App
app = gr.TabbedInterface(
    [register_interface, recognize_interface, download_interface, delete_interface],
    ["Register", "Recognize", "Download Attendance", "Delete Faces"]
)

# Launch Gradio app
app.launch()

	import os
	import pandas as pd
	from datetime import datetime
	from deepface import DeepFace
	import gradio as gr
	from PIL import Image

	# Ensure necessary directories exist
	os.makedirs("registered_faces", exist_ok=True)
	os.makedirs("attendance", exist_ok=True)

	# Adjust the similarity threshold
	SIMILARITY_THRESHOLD = 0.4 # Lower means stricter; increase for more flexibility

	# 1. Register a face
	def register_face(image, name):
	if not name.strip():
	return "Please provide a valid name."

	file_path = f"registered_faces/{name}.jpg"
	image.save(file_path)
	return f"Face registered for {name}!"

	# 2. Recognize a face and mark attendance
	def recognize_face(image):
	if not os.listdir("registered_faces"):
	return "No registered faces found. Please register a face first."

	recognized_name = "Unknown"
	debug_logs = [] # Store debug information

	# Save the uploaded image temporarily
	temp_image_path = "temp_recognition.jpg"
	image.save(temp_image_path)

	for registered_file in os.listdir("registered_faces"):
	registered_path = f"registered_faces/{registered_file}"
	try:
	# Perform face verification
	result = DeepFace.verify(
	img1_path=registered_path,
	img2_path=temp_image_path,
	enforce_detection=True, # Ensure face detection
	model_name="Facenet", # Use a different model for testing
	distance_metric="cosine" # Use cosine distance for better comparison
	)
	similarity = result["distance"]
	debug_logs.append(f"Compared with {registered_file}: Verified={result['verified']}, Distance={similarity}")

	# If a match is found
	if similarity < SIMILARITY_THRESHOLD:
	recognized_name = os.path.splitext(registered_file)[0]
	break
	except Exception as e:
	debug_logs.append(f"Error comparing with {registered_file}: {str(e)}")
	continue

	# Remove temporary file after processing
	if os.path.exists(temp_image_path):
	os.remove(temp_image_path)

	# Log attendance if recognized
	if recognized_name != "Unknown":
	now = datetime.now().strftime("%Y-%m-%d %H:%M:%S")
	attendance_file = "attendance/attendance.csv"

	if not os.path.exists(attendance_file):
	df = pd.DataFrame(columns=["Name", "Time"])
	else:
	df = pd.read_csv(attendance_file)

	df = pd.concat([df, pd.DataFrame({"Name": [recognized_name], "Time": [now]})], ignore_index=True)
	df.to_csv(attendance_file, index=False)

	# Debug output
	return f"Recognized: {recognized_name}\nDebug Logs:\n" + "\n".join(debug_logs)

	# 3. Download attendance records
	def download_attendance():
	attendance_file = "attendance/attendance.csv"
	if os.path.exists(attendance_file):
	return attendance_file
	else:
	return "No attendance records found."

	# 4. Delete all registered faces
	def delete_registered_faces():
	folder = "registered_faces"
	if os.listdir(folder):
	for file in os.listdir(folder):
	os.remove(os.path.join(folder, file))
	return "All registered faces have been deleted."
	return "No registered faces to delete."

	# Gradio interfaces
	register_interface = gr.Interface(
	fn=register_face,
	inputs=[gr.Image(type="pil"), gr.Text(label="Name")],
	outputs="text",
	title="Register Face"
	)

	recognize_interface = gr.Interface(
	fn=recognize_face,
	inputs=gr.Image(type="pil"), # Corrected to use "pil"
	outputs="text",
	title="Recognize Face"
	)

	download_interface = gr.Interface(
	fn=download_attendance,
	inputs=None,
	outputs="file",
	title="Download Attendance"
	)

	delete_interface = gr.Interface(
	fn=delete_registered_faces,
	inputs=None,
	outputs="text",
	title="Delete Registered Faces"
	)

	# Tabbed Gradio App
	app = gr.TabbedInterface(
	[register_interface, recognize_interface, download_interface, delete_interface],
	["Register", "Recognize", "Download Attendance", "Delete Faces"]
	)

	# Launch Gradio app
	app.launch()