Flask OpenCV Send and Receive Images in Bytes

user_5270946 0
Viewed 8253

I want to send and receive images in bytes in my flask API. I also want to send some json alongside the image. How can I achieve this?

Below is my current solution that does not work

flask: 

    @app.route('/add_face', methods=['GET', 'POST'])
    def add_face():
    if request.method == 'POST':
        # print(request.json)
        nparr = np.fromstring(request.form['img'], np.uint8)
        print(request.form['img'])
        img = cv2.imdecode(nparr, cv2.IMREAD_COLOR)

        cv2.imshow("frame", img)
        cv2.waitKey(1)

    return "list of names & faces"

client: 

def save_encoding(img_file):
    URL = "http://localhost:5000/add_face"
    img = open(img_file, 'rb').read()

    response = requests.post(URL, data={"name":"obama", "img":str(img)})
    print(response.content)

produced error:

cv2.imshow("frame", img)
cv2.error: OpenCV(3.4.3) /io/opencv/modules/highgui/src/window.cpp:356: error: (-215:Assertion failed) size.width>0 && size.height>0 in function 'imshow'
flask opencv api python
2 Answers

The following has worked for me. I don't have the client code but I have a curl request. That should do the trick,

SERVER

from flask import request
from PIL import Image
import io

@app.route("/add_face", methods=["POST"])
def predict():
    image = request.files["image"]
    image_bytes = Image.open(io.BytesIO(image.read()))

CLIENT SIDE

curl -X POST -F image=@PATH/TO/FILE 'http://localhost:5000/add_face'
user_25470170

It is easier to send images in base64 format, by doing that you get rid of problems about sending/receiving binary data since you just work with a string. Also it is more convenient in web stuff. Tested code below: Server side: 

from flask import Flask, render_template, request
import pandas as pd
import cv2
import numpy as np
import base64

app = Flask(__name__)



@app.route('/add_face', methods=['GET', 'POST'])
def add_face():
    if request.method == 'POST':
        #  read encoded image
        imageString = base64.b64decode(request.form['img'])

        #  convert binary data to numpy array
        nparr = np.fromstring(imageString, np.uint8)

        #  let opencv decode image to correct format
        img = cv2.imdecode(nparr, cv2.IMREAD_ANYCOLOR);
        cv2.imshow("frame", img)
        cv2.waitKey(0)

    return "list of names & faces"
if __name__ == '__main__':
    app.run(debug=True, port=5000)

Client side: 

import requests
import base64

URL = "http://localhost:5000/add_face"

#  first, encode our image with base64
with open("block.png", "rb") as imageFile:
    img = base64.b64encode(imageFile.read())

response = requests.post(URL, data={"name":"obama", "img":str(img)})
print(response.content)

You can use COLOR instead of ANYCOLOR if you are sure about your input images.

user_76025510
Related