-
Notifications
You must be signed in to change notification settings - Fork 6
/
rest_service.py
50 lines (36 loc) · 1.38 KB
/
rest_service.py
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
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
import os
import numpy as np
from flask_cors import cross_origin
from flask import Flask, jsonify, request
from keras.models import load_model
from utils.dice import dice_coef
from utils.file_utils import convert_to_base64, file_upload
from PIL import Image
IMG_COLS = 128
IMG_ROWS = 128
UPLOAD_PATH = '/Users/akararg/Desktop'
model = load_model('/Users/akararg/Desktop/unet_model_16_filters_kernel10x10.h5',custom_objects={'dice_coef':dice_coef})
# setting config file location
app = Flask(__name__)
@app.route('/segment/cxr', methods=['POST'])
@cross_origin()
def segment_cxr():
parsed_files = request.files
# get file for upload
filename = file_upload(parsed_files, UPLOAD_PATH)
if filename == None:
output = {"Error": "A file has not been sent to the server!"}
return jsonify(output)
image = Image.open(os.path.join(UPLOAD_PATH, filename))
# convert to grayscale and reshape the array to the network input
image = image.resize((IMG_COLS, IMG_ROWS))
image = image.convert('L')
image = np.asarray(image, dtype=np.float32)
image/=255
image = np.reshape(image,(1,IMG_COLS, IMG_ROWS, 1))
segmentation = model.predict(image).argmax(axis=3)
segmentation = np.squeeze(segmentation)
image_encoded = convert_to_base64(segmentation)
return image_encoded
if __name__ == '__main__':
app.run(host='0.0.0.0', port=5201, debug=False)