app.py

import gradio as gr
from deepface import DeepFace
import easyocr
import matplotlib.pyplot as plt
from PIL import Image
import os

def convert_to_jpg(image_path):
    try:
        img = Image.open(image_path)
        jpg_path = os.path.splitext(image_path)[0] + '.jpg'
        img.convert('RGB').save(jpg_path, 'JPEG')
        return jpg_path
    except Exception as e:
        raise ValueError(f"Error converting image: {str(e)}")

def perform_handwriting_recognition_easyocr(image_path):
    try:
        jpg_path = convert_to_jpg(image_path)
        reader = easyocr.Reader(['en', 'hi'], gpu=False)
        results = reader.readtext(jpg_path)
        if not results:
            return "No text found."
        recognized_texts = [result[1] for result in results]
        return " ".join(recognized_texts)
    except Exception as e:
        return f"Error in handwriting recognition: {str(e)}"

def analyze_face(image_path):
    try:
        jpg_path = convert_to_jpg(image_path)
        handwriting_result = perform_handwriting_recognition_easyocr(jpg_path)
        analysis = DeepFace.analyze(img_path=jpg_path, actions=["age", "gender", "emotion", "race"], enforce_detection=False)
        
        # Handle multiple faces in the image
        if isinstance(analysis, list):
            analysis = analysis[0]  # Use the first detected face
        
        gender_scores = analysis.get("gender", {"Man": 0, "Woman": 0})
        predicted_gender = "Male" if gender_scores.get("Man", 0) > gender_scores.get("Woman", 0) else "Female"
        
        outputs = {
            "age": analysis.get("age", "Unknown"),
            "gender_scores": gender_scores,
            "predicted_gender": predicted_gender,
            "emotion": analysis.get("dominant_emotion", "Unknown"),
            "race": analysis.get("dominant_race", "Unknown"),
            "handwriting": handwriting_result,
        }
        return outputs
    except Exception as e:
        return {
            "age": "Error",
            "gender_scores": {"Man": 0, "Woman": 0},
            "predicted_gender": "Error",
            "emotion": "Error",
            "race": "Error",
            "handwriting": f"Error: {str(e)}"
        }

def verify_faces(image_path1, image_path2):
    try:
        jpg_path1 = convert_to_jpg(image_path1)
        jpg_path2 = convert_to_jpg(image_path2)
        result = DeepFace.verify(img1_path=jpg_path1, img2_path=jpg_path2, enforce_detection=False)
        confidence = result["distance"]
        verified = result["verified"]
        return f"Match Confidence: {confidence:.2f}", f"Are they the same person? {'Yes' if verified else 'No'}"
    except Exception as e:
        return f"Error: {str(e)}", ""

def visualize_gender(gender_scores):
    labels = ["Male", "Female"]
    scores = [gender_scores.get("Man", 0), gender_scores.get("Woman", 0)]
    fig, ax = plt.subplots(figsize=(4, 3))
    ax.bar(labels, scores, color=["blue", "pink"])
    ax.set_ylabel("Confidence (%)")
    ax.set_title("Gender Prediction")
    plt.tight_layout()
    plt.close(fig)
    return fig

def display_results(image_path):
    results = analyze_face(image_path)
    gender_chart = visualize_gender(results["gender_scores"])
    return results["age"], gender_chart, results["predicted_gender"], results["emotion"], results["race"], results["handwriting"]

iface = gr.Blocks()

with iface:
    with gr.Tab("Facial and Handwriting Analysis"):
        with gr.Row():
            image_input = gr.Image(type="filepath", label="Upload Image")
        with gr.Row():
            age_output = gr.Textbox(label="Predicted Age")
            gender_plot = gr.Plot(label="Gender Confidence")
            predicted_gender_output = gr.Textbox(label="Predicted Gender")
        with gr.Row():
            emotion_output = gr.Textbox(label="Predicted Emotion")
            race_output = gr.Textbox(label="Predicted Race")
            handwriting_output = gr.Textbox(label="Handwriting Recognition")
        submit_button = gr.Button("Analyze")
        submit_button.click(
            display_results,
            inputs=image_input,
            outputs=[
                age_output, gender_plot, predicted_gender_output,
                emotion_output, race_output, handwriting_output
            ]
        )
    with gr.Tab("Face Verification"):
        with gr.Row():
            image_input1 = gr.Image(type="filepath", label="Upload First Image")
            image_input2 = gr.Image(type="filepath", label="Upload Second Image")
        with gr.Row():
            match_confidence = gr.Textbox(label="Match Confidence")
            verification_result = gr.Textbox(label="Verification Result")
        verify_button = gr.Button("Verify")
        verify_button.click(
            verify_faces,
            inputs=[image_input1, image_input2],
            outputs=[match_confidence, verification_result]
        )

iface.launch()