perform_classification_in_server.py

#!/usr/bin/env python3
# -*- coding: utf-8 -*-
#----------------------------------------------------------------------------
# Created By  : Wanying Li
# Created Date: May 15, 2017
# ---------------------------------------------------------------------------
"""
The server receives the eardrum image from Raspberry, 
then classifies this image using the classification model we trained and saved (which is named CNN_clf).
The server sends the results back to Raspberry.
""" 
# ---------------------------------------------------------------------------

import pickle
import numpy as np
import sklearn
import tensorflow as tf
import tensorflow.python.platform
from tensorflow.python.platform import gfile


def create_graph(model_path):
    """
    The create_graph function loads the inception model to memory. This function should be called before
    calling extract_features or extract_features_single_img.
    
    Input:
        model_path = path to inception model in protobuf form.
    """
    with gfile.FastGFile(model_path, 'rb') as f:
        graph_def = tf.GraphDef()
        graph_def.ParseFromString(f.read())
        _ = tf.import_graph_def(graph_def, name='')


def extract_features_single_img(image_path, verbose=False):
    """
    The extract_features_single_img function computes the inception bottleneck feature for one single image.
    
    Input:
        image_path = directory path of the image
    Output: 
        feature = 2-d np array in the shape of (1, 2048)
    """
    feature_dimension = 2048
    feature = np.empty((1, feature_dimension))

    with tf.Session() as sess:
        flattened_tensor = sess.graph.get_tensor_by_name('pool_3:0')
 
        if verbose:
            print('Processing %s...' % (image_path))

        if not gfile.Exists(image_path):
            tf.logging.fatal('File does not exist %s', image)

        image_data = gfile.FastGFile(image_path,'rb').read()
        feature_tmp = sess.run(flattened_tensor, {'DecodeJpeg/contents:0': image_data})
        feature[:] = np.squeeze(feature_tmp)

    return feature


# create a graph from the Inception V3.0 model
model_path = '/inception_dec_2015/tensorflow_inception_graph.pb'
create_graph(model_path)

# load classifier
clf = pickle.load(open('CNN_clf','rb'))

# setup webserver
UPLOAD_FOLDER = 'uploaded_image'
app = Flask(__name__)
app.config['UPLOAD_FOLDER'] = UPLOAD_FOLDER


@app.route('/classification', methods=['POST'])
def classification():
    # upload user's photo to webserver and save the photo to the UPLOAD_FOLDER on the webserver
    file = request.files['file']
    filepath = os.path.join(UPLOAD_FOLDER,'data.JPG')
    file.save(filepath)
    
    # extraction feature of an user input image
    img_input = filepath
    img_feature = extract_features_single_img(img_input)
    
    # perform image classification
    y_pred = clf.predict(img_feature)
    output = np.array2string(y_pred[0])
    print(output)

    return output

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