File size: 1,117 Bytes
0c66baa d3e1c3a 0c66baa d3e1c3a 0c66baa d3e1c3a 0c66baa d3e1c3a 0c66baa d3e1c3a 0c66baa d3e1c3a | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 | import cv2
import numpy as np
from PIL import Image
import io
from mtcnn import MTCNN
import gradio as gr
detector = MTCNN()
def detect_faces_gradio(image):
img_rgb = cv2.cvtColor(np.array(image), cv2.COLOR_RGB2BGR)
faces = detector.detect_faces(img_rgb)
processed_img = np.array(image).copy()
if faces:
for face in faces:
bounding_box = face['box']
keypoints = face['keypoints']
cv2.rectangle(processed_img,
(bounding_box[0], bounding_box[1]),
(bounding_box[0] + bounding_box[2], bounding_box[1] + bounding_box[3]),
(0, 255, 0),
2)
for key, value in keypoints.items():
cv2.circle(processed_img, (int(value[0]), int(value[1])), 2, (0, 0, 255), 2)
return processed_img
iface = gr.Interface(fn=detect_faces_gradio, inputs=gr.Image(type="pil"), outputs="image",
title="MTCNN Face Detection",
description="Upload an image and see the detected faces and landmarks.")
iface.launch() |