Styjun

I am updating an image sent through post method using cv2 in the flask backend.
Then when I am trying to save that image using cv2.imwrite() to a path inside the flask server and then send that path as image-src response back,its unable to find the file.I have also tried to convert the opencv version image to pillow based image and then save it to the flask server,but also failed.I mainly want to know how can I save the image using cv2/pillow inside the flask backend and serve that url as image source to render in the front-end.
My code is below:

# import the necessary packages
from PIL import Image
import numpy as np
import flask
import io
import base64
from flask import request, render_template, Flask
import cv2
from keras.models import model_from_json
import re, os
import uuid

# initialize our Flask application and the Keras model
model = None
full_path = None

def create_app():
 app = Flask(__name__)
 APP_ROOT = os.path.dirname(os.path.abspath(__file__))
 app.config["UPLOAD_FOLDER"]=os.path.join(APP_ROOT,"uploads")
 def load_model():
 print("hello")
 global model
 # load json and create model
 json_file = open('model.json', 'r')
 loaded_model_json = json_file.read()
 json_file.close()
 model = model_from_json(loaded_model_json)
 # load weights into new model
 model.load_weights("model.h5")
 print("Loaded model from disk")
 model.compile(loss='categorical_crossentropy',
 optimizer="rmsprop",
 metrics=['accuracy'])

 load_model()
 return app


app = create_app()

def do_prediction(im):
 global full_path
 label_map = ["Angry", "Fear", "Happy",
 "Sad", "Surprise", "Neutral"]
 gray = cv2.cvtColor(im, cv2.COLOR_BGR2GRAY)
 print(im)
 face_cascade = cv2.CascadeClassifier("cascade_frontface.xml")
 faces = face_cascade.detectMultiScale(gray, 1.2, 3, minSize=(80, 80))
 print(len(faces))
 font = cv2.FONT_HERSHEY_SIMPLEX
 for (x, y, w, h) in faces:
 cv2.rectangle(im, (x, y), (x + w, y + h), (0, 255, 0), 2, 5)
 face_crop = im[y:y + h, x:x + w]
 face_crop = cv2.resize(face_crop, (48, 48))
 face_crop = cv2.cvtColor(face_crop, cv2.COLOR_BGR2GRAY)
 face_crop = face_crop.astype('float32') / 255.0
 face_crop = np.asarray(face_crop)
 face_crop = face_crop.reshape(
 1, face_crop.shape[0], face_crop.shape[1], 1)
 result = label_map[np.argmax(model.predict(face_crop))]
 cv2.putText(im, result, (x, y), font, 2, (200, 0, 0), 3, cv2.LINE_AA)

 random_value = str(uuid.uuid4())
 new_image = random_value + ".jpg"
 if not os.path.isdir(app.config["UPLOAD_FOLDER"]):
 os.mkdir(app.config["UPLOAD_FOLDER"])
 full_path = os.path.join(app.config['UPLOAD_FOLDER'], new_image)
 # cv2.imshow("image",im)
 # cv2.imwrite(full_path, im)
 # predicted_image = cv2.cvtColor(im, cv2.COLOR_BGR2RGB)
 pil_image=cv2.cvtColor(im, cv2.COLOR_BGR2RGB)
 updated_image = Image.fromarray(pil_image, mode='RGB')
 print("converted to pil")
 print(updated_image)
 updated_image.save(full_path,'JPEG')

@app.route("/", methods=["GET"])
def index():
 return render_template("predict.html")


@app.route("/predict", methods=["POST"])
def predict():
 # initialize the data dictionary that will be returned from the
 # view
 data = "success": False
 message = request.get_json(force=True)
 encoded = message["image"]
 encoded = re.sub('^data:image/.+;base64,', '', encoded)
 decoded = base64.b64decode(encoded)
 image = Image.open(io.BytesIO(decoded))
 open_cv_image = cv2.cvtColor(np.array(image),cv2.COLOR_RGB2BGR)
 # # Convert RGB to BGR
 # open_cv_image = open_cv_image[:, :, ::-1].copy()
 print(open_cv_image)
 # classify the input image and then initialize the list
 # of predictions to return to the client

 do_prediction(open_cv_image)
 # indicate that the request was a success
 data["success"] = True
 data["image"] = full_path

 # return the data dictionary as a JSON response
 return flask.jsonify(data)


if __name__ == "__main__":
 app.run()