# -*- coding: utf-8 -*-from __future__ import division, print_function, absolute_importimport os
import numpy as np
import pickle
import tflearnfrom PIL import Image
from tflearn.layers.core import input_data, dropout, fully_connected
from tflearn.layers.conv import conv_2d, max_pool_2d, avg_pool_2d
from tflearn.layers.merge_ops import merge
from tflearn.layers.estimator import regression
from tflearn.data_utils import to_categorical, shuffle
from sklearn.model_selection import train_test_split
from sklearn.metrics import classification_report, confusion_matrix
from tflearn.layers.conv import highway_conv_2d, max_pool_2d
from tflearn.layers.normalization import local_response_normalization, batch_normalizationdef resize_image(in_image, new_width, new_height, out_image=None,resize_mode=Image.ANTIALIAS):""" Resize an image.Arguments:in_image: `PIL.Image`. The image to resize.new_width: `int`. The image new width.new_height: `int`. The image new height.out_image: `str`. If specified, save the image to the given path.resize_mode: `PIL.Image.mode`. The resizing mode.Returns:`PIL.Image`. The resize image."""img = in_image.resize((new_width, new_height), resize_mode)if out_image:img.save(out_image)return imgdef convert_color(in_image, mode):""" Convert image color with provided `mode`. """return in_image.convert(mode)def pil_to_nparray(pil_image):""" Convert a PIL.Image to numpy array. """pil_image.load()return np.asarray(pil_image, dtype="float32")def iterbrowse(path):for home, dirs, files in os.walk(path):for filename in files:yield os.path.join(home, filename)def directory_to_samples(directory, flags):""" Read a directory, and list all subdirectories files as class sample """samples = []targets = []# label class is from 0 !!!label = 0try:  # Python 2classes = sorted(os.walk(directory).next()[1])except Exception:  # Python 3classes = sorted(os.walk(directory).__next__()[1])for c in classes:c_dir = os.path.join(directory, c)try:  # Python 2walk = os.walk(c_dir).next()except Exception:  # Python 3walk = os.walk(c_dir).__next__()for sample in walk[2]:if any(flag in sample for flag in flags):samples.append(os.path.join(c_dir, sample))targets.append(label)label += 1return samples, targets# Get the pixel from the given image
def get_pixel(image, i, j):# Inside image bounds?width, height = image.sizeif i > width or j > height:return None# Get Pixelpixel = image.getpixel((i, j))return pixel# Create a Grayscale version of the image
def convert_to_one_channel(image):# !!! I assume that the png file is grayscale. And R == G == B !!!! So I check it..."""for i in range(len(image)):for j in range(len(image[i])):pixel = image[i][j]# Get R, G, B values (This are int from 0 to 255)assert len(pixel) == 3red = pixel[0]green = pixel[1]blue = pixel[2]assert red == green == blueassert 0 <= red <= 1"""# Just extract 1 channel datareturn image[:, :, [0]]def image_dirs_to_samples(directory, resize=None, convert_gray=False,filetypes=None):print("Starting to parse images...")samples, targets = directory_to_samples(directory, flags=filetypes)for i, s in enumerate(samples):print("Process %d th file %s" % (i+1, s))samples[i] = Image.open(s)  # Load an image, returns PIL.Image.if resize:######################## TODO #######################samples[i] = resize_image(samples[i], resize[0],resize[1])######################### TODO ####################### convert to more data# if convert_gray:#    samples[i] = convert_color(samples[i], 'L')samples[i] = pil_to_nparray(samples[i])samples[i] /= 255.  # ormalize a list of sample image data in the range of 0 to 1samples[i] = convert_to_one_channel(samples[i]) # just want 1 channel dataprint("Parsing Done!")return samples, targetsdef load_data(dirname, resize_pics=(128, 128), shuffle_data=True):dataset_file = os.path.join(dirname, 'data.pkl')try:X, Y, org_labels = pickle.load(open(dataset_file, 'rb'))except Exception:# X, Y = image_dirs_to_samples(os.path.join(dirname, 'train/'), resize_pics, False, ['.jpg', '.png'])X, Y = image_dirs_to_samples(dirname, resize_pics, False,['.jpg', '.png'])  # TODO, memory is too small to load all dataorg_labels = YY = to_categorical(Y, np.max(Y) + 1)  # First class is '0', Convert class vector (integers from 0 to nb_classes)if shuffle_data:X, Y, org_labels = shuffle(X, Y, org_labels)pickle.dump((X, Y, org_labels), open(dataset_file, 'wb'))return X, Y, org_labelsclass EarlyStoppingCallback(tflearn.callbacks.Callback):def __init__(self, val_acc_thresh):# Store a validation accuracy threshold, which we can compare against# the current validation accuracy at, say, each epoch, each batch step, etc.self.val_acc_thresh = val_acc_threshdef on_epoch_end(self, training_state):"""This is the final method called in trainer.py in the epoch loop.We can stop training and leave without losing any information with a simple exception."""# print dir(training_state)print("Terminating training at the end of epoch", training_state.epoch)if training_state.val_acc >= self.val_acc_thresh and training_state.acc_value >= self.val_acc_thresh:raise StopIterationdef on_train_end(self, training_state):"""Furthermore, tflearn will then immediately call this method after we terminate training,(or when training ends regardless). This would be a good time to store any additionalinformation that tflearn doesn't store already."""print("Successfully left training! Final model accuracy:", training_state.acc_value)def get_model(width, height, classes=40):# TODO, modify model# Real-time data preprocessingimg_prep = tflearn.ImagePreprocessing()# Real-time data preprocessingimg_prep = tflearn.ImagePreprocessing()img_prep.add_featurewise_zero_center(per_channel=True)img_prep.add_featurewise_stdnorm()network = input_data(shape=[None, width, height, 1], data_preprocessing=img_prep)  # if RGB, 224,224,3network = conv_2d(network, 32, 3, activation='relu')network = max_pool_2d(network, 2)network = conv_2d(network, 64, 3, activation='relu')network = conv_2d(network, 64, 3, activation='relu')network = max_pool_2d(network, 2)network = fully_connected(network, 512, activation='relu')network = dropout(network, 0.5)network = fully_connected(network, classes, activation='softmax')network = regression(network, optimizer='adam',loss='categorical_crossentropy',learning_rate=0.001)model = tflearn.DNN(network, tensorboard_verbose=0)return modelif __name__ == "__main__":width, height = 32, 32X, Y, org_labels = load_data(dirname="data", resize_pics=(width, height))trainX, testX, trainY, testY = train_test_split(X, Y, test_size=0.2, random_state=666)print("sample data:")print(trainX[0])print(trainY[0])print(testX[-1])print(testY[-1])model = get_model(width, height, classes=100)filename = 'cnn_handwrite-acc0.8.tflearn'# try to load model and resume training#try:#    model.load(filename)#    print("Model loaded OK. Resume training!")#except:#    pass# Initialize our callback with desired accuracy threshold.early_stopping_cb = EarlyStoppingCallback(val_acc_thresh=0.9)try:model.fit(trainX, trainY, validation_set=(testX, testY), n_epoch=500, shuffle=True,snapshot_epoch=True, # Snapshot (save & evaluate) model every epoch.show_metric=True, batch_size=32, callbacks=early_stopping_cb, run_id='cnn_handwrite')except StopIteration as e:print("OK, stop iterate!Good!")model.save(filename)# predict all data and calculate confusion_matrixmodel.load(filename)pro_arr =model.predict(X)predict_labels = np.argmax(pro_arr, axis=1)print(classification_report(org_labels, predict_labels))print(confusion_matrix(org_labels, predict_labels))

---------------------------------
Run id: cnn_handwrite
Log directory: /tmp/tflearn_logs/
---------------------------------
Preprocessing... Calculating mean over all dataset (this may take long)...
Mean: [ 0.89235026] (To avoid repetitive computation, add it to argument 'mean' of `add_featurewise_zero_center`)
---------------------------------
Preprocessing... Calculating std over all dataset (this may take long)...
STD: 0.192279 (To avoid repetitive computation, add it to argument 'std' of `add_featurewise_stdnorm`)
---------------------------------
Training samples: 19094
Validation samples: 4774
--
Training Step: 597  | total loss: 0.70288 | time: 40.959ss
| Adam | epoch: 001 | loss: 0.70288 - acc: 0.7922 | val_loss: 0.54380 - val_acc: 0.8460 -- iter: 19094/19094
--
Terminating training at the end of epoch 1Training Step: 1194  | total loss: 0.48860 | time: 40.245s
| Adam | epoch: 002 | loss: 0.48860 - acc: 0.8783 | val_loss: 0.37020 - val_acc: 0.8923 -- iter: 19094/19094
--
Terminating training at the end of epoch 2
Training Step: 1791  | total loss: 0.35790 | time: 41.315ss
| Adam | epoch: 003 | loss: 0.35790 - acc: 0.9090 | val_loss: 0.34719 - val_acc: 0.9049 -- iter: 19094/19094
--
Terminating training at the end of epoch 3
Successfully left training! Final model accuracy: 0.908959209919
OK, stop iterate!Good!precision    recall  f1-score   support0       1.00      0.99      0.99       2391       0.95      0.96      0.96       2372       0.91      0.98      0.94       2403       0.90      0.98      0.94       2394       0.96      0.98      0.97       2395       0.94      0.97      0.96       2396       0.98      0.98      0.98       2397       0.84      0.99      0.91       2408       0.99      0.87      0.93       2399       0.95      0.98      0.96       23910       0.97      0.94      0.96       24011       0.95      0.98      0.97       24012       0.92      0.99      0.95       24013       0.95      0.96      0.96       23914       0.94      0.94      0.94       23615       0.94      0.97      0.96       24016       0.94      0.98      0.96       24017       0.97      0.99      0.98       24018       0.94      0.93      0.94       24019       1.00      0.95      0.98       23920       0.96      0.98      0.97       24021       0.98      0.91      0.95       23922       0.97      0.95      0.96       23923       1.00      0.97      0.98       23924       0.94      0.98      0.96       24025       0.98      0.98      0.98       23726       0.91      1.00      0.95       23927       0.91      0.96      0.93       23928       0.97      0.88      0.92       23929       1.00      0.98      0.99       24030       0.99      0.94      0.96       23931       0.97      0.97      0.97       23732       0.94      0.98      0.96       23633       0.94      0.96      0.95       23934       0.98      0.99      0.98       23935       0.99      0.98      0.99       24036       0.96      0.92      0.94       23937       1.00      0.93      0.96       24038       0.96      0.99      0.98       23839       0.98      0.97      0.97       23840       0.92      0.90      0.91       24041       0.96      0.97      0.96       23742       0.98      0.97      0.97       24043       0.95      0.96      0.95       23944       0.97      0.96      0.97       23945       0.95      0.94      0.95       23946       0.93      0.96      0.94       23247       0.98      0.91      0.94       23748       0.95      0.97      0.96       23949       0.97      0.80      0.88       22650       0.90      0.95      0.92       24051       0.98      0.99      0.99       23652       0.96      0.90      0.93       24053       0.99      0.96      0.97       23554       0.97      0.93      0.95       24055       0.99      0.98      0.99       24056       0.97      0.92      0.95       23957       0.97      0.97      0.97       23958       1.00      0.98      0.99       23859       0.92      0.98      0.95       24060       0.99      0.90      0.94       24061       1.00      0.99      0.99       23862       0.92      0.95      0.94       23963       0.92      0.98      0.95       23864       0.98      0.92      0.95       24065       0.99      0.92      0.95       23966       0.98      0.99      0.99       24067       0.95      0.95      0.95       24068       0.96      0.98      0.97       23969       0.97      0.97      0.97       23970       0.98      0.94      0.96       24071       0.91      0.96      0.93       23972       0.98      0.97      0.97       23973       0.99      0.89      0.94       24074       0.97      0.99      0.98       23775       0.89      0.97      0.92       24076       0.97      0.96      0.97       24177       0.89      0.91      0.90       24078       1.00      0.89      0.94       23979       0.90      0.98      0.94       23980       0.89      0.96      0.92       24081       0.96      0.71      0.81       22582       0.95      1.00      0.97       23883       0.67      0.96      0.79       23984       0.97      0.85      0.91       24085       0.95      0.98      0.96       23986       0.99      0.93      0.96       24087       0.98      0.91      0.94       23988       0.97      0.97      0.97       24089       0.97      0.94      0.95       23990       0.97      0.96      0.96       23691       0.91      0.97      0.94       23992       0.98      0.95      0.96       24093       0.98      0.97      0.98       23994       0.98      0.95      0.97       24095       0.98      0.99      0.99       23996       0.95      0.97      0.96       24097       0.98      0.97      0.98       23998       0.95      0.98      0.97       23799       0.97      0.96      0.97       239avg / total       0.96      0.95      0.95     23868[[237   0   0 ...,   0   0   0][  0 228   0 ...,   0   0   0][  0   0 235 ...,   0   0   0]..., [  0   0   0 ..., 233   0   0][  0   0   0 ...,   0 233   0][  0   0   0 ...,   0   0 230]]

    model = get_model(width, height, classes=100)filename = 'cnn_handwrite-acc0.8.tflearn'# try to load model and resume training#try:#    model.load(filename)#    print("Model loaded OK. Resume training!")#except:#    pass# Initialize our callback with desired accuracy threshold.early_stopping_cb = EarlyStoppingCallback(val_acc_thresh=0.8)try:model.fit(trainX, trainY, validation_set=(testX, testY), n_epoch=500, shuffle=True,snapshot_epoch=True, # Snapshot (save & evaluate) model every epoch.show_metric=True, batch_size=32, callbacks=early_stopping_cb, run_id='cnn_handwrite')except StopIteration as e:print("OK, stop iterate!Good!")model.save(filename)del tf.get_collection_ref(tf.GraphKeys.TRAIN_OPS)[:]"""# print op namewith tf.Session() as sess:init_op = tf.initialize_all_variables()sess.run(init_op)for v in sess.graph.get_operations():print(v.name)"""filename = 'cnn_handwrite-acc0.8.infer.tflearn'model.save(filename)

def inference(image):print('inference')temp_image = Image.open(image).convert('L')temp_image = temp_image.resize((FLAGS.image_size, FLAGS.image_size), Image.ANTIALIAS)temp_image = np.asarray(temp_image) / 255.0temp_image = temp_image.reshape([-1, 32, 32, 1])from tensorflow.python.platform import gfilewith tf.Graph().as_default():output_graph_def = tf.GraphDef()with open("frozen_model.pb", "rb") as f:output_graph_def.ParseFromString(f.read())tensors = tf.import_graph_def(output_graph_def, name="")#print tensorswith tf.Session() as sess:init = tf.global_variables_initializer()sess.run(init)op = sess.graph.get_operations()"""for m in op:print(m.values())"""op = sess.graph.get_tensor_by_name("FullyConnected_1/Softmax:0")input_tensor = sess.graph.get_tensor_by_name('InputData/X:0')probs = sess.run(op,feed_dict = {input_tensor:temp_image})print probsresult = []for word in probs:result.append(np.argsort(-word)[:3])return resultdef main(_):image_path = './data/test/00098/104405.png'#image_path = '../data/00010/17724.png'final_predict_val = inference(image_path)logger.info('the result info label {0} predict index {1}'.format(98, final_predict_val))

[[  8.42533936e-08   1.60850794e-11   2.60133332e-10   2.42555542e-144.96124599e-08   4.45251297e-15   3.98175590e-11   1.64476592e-117.03968351e-13   5.42319011e-12   8.55469237e-11   4.91866422e-131.77282828e-07   4.05237593e-10   3.13049003e-10   1.34780919e-112.05803235e-06   2.87827305e-07   1.47789994e-12   2.53391891e-113.77086790e-13   2.02639586e-10   9.03167027e-13   3.96698889e-111.30850096e-11   5.71980917e-12   3.03487374e-11   2.04132298e-146.25303683e-13   1.46122332e-07   2.17450633e-07   1.69623715e-096.80857757e-12   2.52643609e-13   6.56771096e-11   8.55152287e-161.34496514e-09   1.22644633e-06   1.12011307e-07   7.93476283e-058.24334611e-12   4.77531155e-14   9.39397757e-13   2.38438267e-142.11416329e-10   5.54395712e-08   2.30046147e-12   2.63584043e-104.70621564e-16   5.14432724e-12   6.42602327e-09   1.62485829e-137.39078274e-08   3.19146315e-12   5.25887156e-09   1.35877786e-131.39127886e-13   2.11998293e-13   9.09501097e-09   9.46486750e-072.47498733e-09   2.74523763e-12   1.02716433e-14   1.02069058e-173.09356682e-16   1.51022904e-15   9.34333665e-13   2.62195051e-143.38079781e-16   7.43019903e-13   1.92409091e-13   3.86611994e-132.61276265e-12   1.07969211e-09   1.30814548e-09   2.44038188e-149.79275905e-13   1.41007803e-10   6.15137758e-12   2.08893070e-101.34751668e-14   2.76824767e-15   7.84100464e-16   7.70873335e-155.45704757e-12   3.69386271e-18   2.06012223e-13   1.62567273e-141.54544960e-03   2.05292008e-06   1.31726174e-09   7.04993663e-094.11338266e-03   3.19344110e-07   3.96519717e-05   2.26919351e-122.39114349e-12   2.35558744e-07   9.94213998e-01   1.10125060e-11]]
the result info label 98 predict index [array([98, 92, 88])]