ageitgey
Capture face in webcam and send to flask
face_recognition version:1.2.3 Python version:3.7.2 Operating System: Window 64bits
I have a question. currently I am working with face recognition project by using face_recognition library provided by Python, flask and OpenCV. Currently my face recognition works well in flask server.
Currently I am working on face capture in flask. What I mean here is when the webcam detect the face it will capture the face and display in flask.
My question is how I can send the capture face in webcam and display in flask? It's great if someone can share the code to me or maybe related tutorial on this.
We are using this way to handle images from Flask:
@self.api.route('/api/v1/demo/face', methods = ['POST'])
def route_face():
img = self.get_img(request)
res = self.module_face.do_process(img)
return self.get_res(res)
def get_img(self, request):
data = request.files['file'].read()
npimg = np.frombuffer(data, np.uint8)
img = cv2.imdecode(npimg, cv2.IMREAD_COLOR)
return img
def get_res(self, res):
resp = Response(response = res, status = 200, mimetype = "application/json")
resp.headers['Content-Type'] = 'application/json; charset=utf-8'
resp.headers['Access-Control-Allow-Origin'] = '*'
return resp
PS: After that, you can use the img as you like.
Post method with cURL:
curl -F "file=@./tests/barbara_palvin.jpg" http://localhost:5000/api/v1/demo/face
I hope it helps, if you are looking for like that, @adamgan919191
Good Luck!
Hye Dentrax, Thank you help and reply :)
Basically I have three file which is app.py camera.py and gallery.html. I attach my code for you reference.
app.py
from flask import Flask, Response, json, render_template
from werkzeug.utils import secure_filename
from flask import request
from os import path, getcwd
import time
import os
app = Flask(__name__)
import cv2
from camera import VideoCamera
app.config['file_allowed'] = ['image/png', 'image/jpeg']
app.config['train_img'] = path.join(getcwd(), 'train_img')
def gen(camera):
while True:
frame = camera.get_frame()
yield (b'--frame\r\n'
b'Content-Type: image/jpeg\r\n\r\n' + frame + b'\r\n\r\n')
@app.route('/video_feed')
def video_feed():
return Response(gen(VideoCamera()),
mimetype='multipart/x-mixed-replace; boundary=frame')
@app.route('/')
def index():
return render_template('index.html')
@app.route('/gallery')
def get_gallery():
images = os.listdir(os.path.join(app.static_folder, "capture_image"))
return render_template('gallery.html', images=images)
app.run()
camera.py
import cv2
import face_recognition
from PIL import Image
import os
import time
dir_path = "C:/tutorial/face_recognition/venv/src4/capture_image"
class VideoCamera(object):
def __init__(self):
self.video = cv2.VideoCapture(0)
def get_frame(self):
success, frame = self.video.read()
small_frame = cv2.resize(frame, (0, 0), fx=0.25, fy=0.25)
rgb_small_frame = small_frame[:, :, ::-1]
face_locations = face_recognition.face_locations(rgb_small_frame,number_of_times_to_upsample=2)
for face_location in face_locations:
top, right, bottom, left = face_location
face_image = rgb_small_frame[top:bottom, left:right]
pil_image = Image.fromarray(face_image)
File_Formatted = ("%s" % (top)) + ".jpg"
file_path = os.path.join( dir_path, File_Formatted)
pil_image.save(file_path)
ret, jpeg = cv2.imencode('.jpg', frame)
return jpeg.tobytes()
gallery.html
<section class="row">
{% for image in images %}
<section class="col-md-4 col-sm-6" style="background-color: green;">
<img src="{{ url_for('static', filename='capture_image/' + image) }}">
</section>
{% endfor %}
</section>
This what i have done so far, the webcam will capture the face in webcam and save in folder. Then send the image to gallery.html. Currently, i am want to display the image real time in html templete without refresh when the image in save folder changing. can you help me on this matter?
Thank you
@adamgan919191 I think that way is a bit slow way. Because you are doing I/O operations. Instead, my advice is to use socket for real-time operations. If you want to get least 60 FPS, do not face recognition in the main loop. Instead, you can run face recognition operations in async, on different thread, each X frame. This can be the fastest solution.
you can run face recognition operations in
async
does face recognition lib support concurrency via async?
