[314dda]: / app.py

Download this file

169 lines (138 with data), 6.5 kB

  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
 48
 49
 50
 51
 52
 53
 54
 55
 56
 57
 58
 59
 60
 61
 62
 63
 64
 65
 66
 67
 68
 69
 70
 71
 72
 73
 74
 75
 76
 77
 78
 79
 80
 81
 82
 83
 84
 85
 86
 87
 88
 89
 90
 91
 92
 93
 94
 95
 96
 97
 98
 99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
"""
=============================================================================================
Project : Chest X-Ray Pathology Detection and Localization using Deep Learning
Author Name : Rammuni Ravidu Suien Silva
UoW No : 16267097
IIT No : 2016134
Module : Final Year Project 20/21
Supervisor : Mr Pumudu Fernando
Prototype : Web Interface - BackEnd [Draft: .v01]
University of Westminster, UK || IIT Sri Lanka
=============================================================================================
"""
import json
import os
from datetime import datetime
import numpy as np
# Flask Imports
from flask import Flask, request, render_template
from flask import send_file
from flask_jsglue import JSGlue
# Tensorflow Keras imports
from tensorflow.keras.models import load_model
# For secure src links
from werkzeug.utils import secure_filename
# System Library import
from lab_cxr_scripts.lab_cxr import CXRPrediction, CXRLocalization
# Model 0, 2 :- xray_labels_set[0] || Model 1 :- xray_labels_set[1]
xray_labels_set = [["Enlarged Cardiomediastinum", "Cardiomegaly", "Lung Lesion", "Lung Opacity", "Edema",
"Consolidation", "Pneumonia", "Atelectasis", "Pneumothorax", "Pleural Effusion",
"Pleural Other", "Fracture", "Support Devices"],
["Nodule", "Cardiomegaly", "Emphysema", "Fibrosis", "Edema", "Consolidation", "Pneumonia",
"Atelectasis", "Pneumothorax", "Pleural Effusion", "Mass", "Infiltration", "Hernia",
"Plueral Thickening"]]
# Labels for classification tasks
xray_labels = xray_labels_set[0]
# Dependency pip install pyopenssl
# Flask Configs
app = Flask(__name__)
app.config['MAX_CONTENT_LENGTH'] = 20 * 1024 * 1024 # Request data limited to 20MB
jsglue = JSGlue(app)
# TODO: USER GUIDE
# model load
models = [[
load_model('models/MIMIC/PAR-64-MODEL-MIMIC-FINAL-2.h5',
custom_objects={'weighted_loss': CXRPrediction.get_weighted_loss(1, 1)}),
load_model('models/MIMIC/PAR-128-MODEL-MIMIC-FINAL-2.h5',
custom_objects={'weighted_loss': CXRPrediction.get_weighted_loss(1, 1)})
], [
load_model('models/NIH/PAR-64-MODEL-FINAL-NIH-2.h5',
custom_objects={'weighted_loss': CXRPrediction.get_weighted_loss(1, 1)}),
load_model('models/NIH/PAR-128-MODEL-FINAL-NIH-2.h5',
custom_objects={'weighted_loss': CXRPrediction.get_weighted_loss(1, 1)})
]]
model = models[0]
cur_cxr_hash = 'none'
"""
==================================================================================================================
Web request functions
==================================================================================================================
"""
# Web page startup
@app.route('/')
def start_web():
return render_template("index.html")
# CXR Image upload API
@app.route('/predict/<int:model_id>', methods=['GET', 'POST'])
def upload(model_id):
if request.method == 'POST':
print("Model ID", model_id)
# Selecting Model and labels set
global model, xray_labels
model = models[model_id % len(models)]
xray_labels = xray_labels_set[model_id % len(xray_labels_set)]
global cur_cxr_hash
preds = []
file_count = len(request.files)
if file_count > 8:
return
for file_num in range(file_count):
# Getting image file from post request through the Web
cxr_img_file = request.files['file_' + str(file_num)]
# Generating Hash of the image file
hashed_filename = CXRPrediction.hash_cxr(cxr_img_file)
print(hashed_filename)
cur_cxr_hash = hashed_filename
# Saving the CXR image to uploads
cxr_img_path = os.path.dirname(__file__)
file_path = os.path.join(
cxr_img_path, 'uploads', secure_filename(hashed_filename))
cxr_img_file.save(file_path)
# Detection results calculation
preds.append(np.array(CXRPrediction.model_predict(file_path, model)[0]).tolist())
# Final results calculation considering the results of all the uploaded images
final_preds = np.round(np.multiply(np.mean(preds, axis=0), 100), 2)
final_preds_max = np.round(np.multiply(np.max(preds, axis=0), 100), 2)
final_preds_min = np.round(np.multiply(np.min(preds, axis=0), 100), 2)
print(final_preds)
# Creating the detection results dictionary/ JSON
predictions_dict = {}
for i in range(0, len(xray_labels)):
det_rate_str = str(final_preds[i]) + "% (" + str(final_preds_max[i]) + "% - " + str(
final_preds_min[i]) + "%)"
predictions_dict[xray_labels[i]] = det_rate_str
# Creating detection result JSON to be sent
json_predictions = json.dumps(predictions_dict, indent=4)
result = json_predictions
return result
return None
@app.route('/localize')
def localization(): # Localization API
global cur_cxr_hash
start = datetime.now()
filepath = 'localizations/' + cur_cxr_hash.split('.')[0]
if os.path.exists(filepath):
file_count = len([name for name in os.listdir(filepath) if os.path.isfile(os.path.join(filepath, name))])
if not file_count == len(xray_labels):
# If the localized img is already there no need to re-process
CXRLocalization.create_cxr_localization_heatmap(cur_cxr_hash, model[len(model) - 1], xray_labels)
else:
# Calling Localization Function
CXRLocalization.create_cxr_localization_heatmap(cur_cxr_hash, model[len(model) - 1], xray_labels)
print(datetime.now() - start)
return str(len(xray_labels)) # Returning the localized labels
# Function for sending the localized CXR image
@app.route('/get_cxr_detect_img/<int:pathology_id>')
def get_cxr_detect_img(pathology_id):
print(pathology_id)
global cur_cxr_hash
localized_image_name = xray_labels[pathology_id] + '-localizedHeatmap-' + cur_cxr_hash
filepath = 'localizations/' + cur_cxr_hash.split('.')[0] + '/'
return send_file(filepath + localized_image_name, mimetype='image/jpg')
# Function for getting symptoms
@app.route('/get_symptoms')
def get_symptoms():
return send_file('static/files/Symptoms.json', mimetype='application/json')
print("Server Running...")
if __name__ == '__main__':
app.run(debug=True) # Debugging