20161119-meganeco - Speaker Deck

Slide 1

Slide 1 text

C-LIS CO., LTD. 5FOTPS'MPXͰझຯͷը૾ऩूαʔόʔΛ࡞Δ݄߸ BCDB

Slide 2

Slide 2 text

C-LIS CO., LTD. ༗ࢁܓೋʢ,FJKJ"3*:"."ʣ $-*4$0 -5% "OESPJEΞϓϦ։ൃνϣοτσΩϧ ػցֶशॳ৺ऀ Photo : Koji MORIGUCHI (AUN CREATIVE FIRM) ΍ͬͯ·ͤΜ

Slide 3

Slide 3 text

݄೔ʙ https://devfestkansai.connpass.com/event/42864/

Slide 4

Slide 4 text

C-LIS CO., LTD. (%(,PCFʢʣ (%(,ZPUPʢʣ ೔ຊ"OESPJEͷձ"OESPJE#B[BBSBOE$POGFSFODFʢʣ (%(,PCFʢ ڞ࠵($16#0TBLBʣ 5FOTPS'MPXษڧձʢ̏ʣʢ ڞ࠵೔ຊ"OESPJEͷձʣ 5FOTPS'MPXษڧձʢ̐ʣʢʣ ͳʹΘ5&$)ಓʢʣ 5FOTPS'MPXษڧձʢʣʢʣ ೔ຊ"OESPJEͷձ"OESPJE#B[BBSBOE$POGFSFODFʢʣ ˡ/FXʂ

Slide 5

Slide 5 text

10% 1SJOU0O%FNBOE ిࢠॻ੶ Impress R&D͔Βൃചத http://amzn.to/2axRog

Slide 6

Slide 6 text

C-LIS CO., LTD.

Slide 7

Slide 7 text

޷Έͷʰ؟ڸ່ͬʱը૾Λ ࣗಈͰऩू͍ͨ͠

Slide 8

Slide 8 text

4PGUXBSF%FTJHO೥݄߸ʰ"OESPJE4UVEJPͷελΠϧͰޮ཰ΞοϓʱΑΓ © ࠜઇΕ͍ ؟ ڸ ͬ ່

Slide 9

Slide 9 text

̎Ϋϥε෼ྨ

Slide 10

Slide 10 text

؟ڸ່ͬσʔληοτ

Slide 11

Slide 11 text

؟ڸ່ͬը૾ຕ ඇ؟ڸ່ͬຕ ܭɿ ຕ ೥݄ ؟ڸ່ͬ ඇ؟ڸ່ͬ ޡݕग़ ؟ڸ່ͬ ඇ؟ڸ່ͬ

Slide 12

Slide 12 text

؟ڸ່ͬը૾ ຕ ඇ؟ڸ່ͬ ຕ ܭɿ ຕ ೥݄ ؟ڸ່ͬ ඇ؟ڸ່ͬ

Slide 13

Slide 13 text

؟ڸ່ͬը૾ ຕ ඇ؟ڸ່ͬ ຕ ܭɿ ຕ ೥݄ ؟ڸ່ͬ ඇ؟ڸ່ͬ

Slide 14

Slide 14 text

ݱࡏͷߏ੒ Downloader Face Detection Megane Detection API TensorFlow ֬ೝɾमਖ਼ ೝࣝ݁Ռ JSON ܇࿅ λΠϜϥΠϯ ϝσΟΞ σʔληοτ ܇࿅

Slide 15

Slide 15 text

%PXOMPBEFS Downloader Face Detection Megane Detection API TensorFlow ֬ೝɾमਖ਼ ೝࣝ݁Ռ JSON ܇࿅ λΠϜϥΠϯ ϝσΟΞ σʔληοτ ܇࿅

Slide 16

Slide 16 text

w͋Β͔͡Ίొ࿥ͨ͠5XJUUFS*%ͷ౤ߘ͢ΔϝσΟΞΛμ΢ϯϩʔυ w5XJUUFS*%ͱը૾63-ΛσʔλϕʔεͰ؅ཧ͢Δ͜ͱͰɺ  ը૾ͷݖརऀΛ֬ೝՄೳʹ %PXOMPBEFS

Slide 17

Slide 17 text

wը૾ͷॏෳ͕සൟʹൃੜͯ͠͠·͏ɻ ఆظతͳ޿ࠂπΠʔτ ιʔγϟϧήʔϜͷϓϨΠը໘ େચɹͳͲ ՝୊

Slide 18

Slide 18 text

JNBHF)BTI 67d85827b2396276 = = =

Slide 19

Slide 19 text

'BDF%FUFDUJPO Downloader Face Detection Megane Detection API TensorFlow ֬ೝɾमਖ਼ ೝࣝ݁Ռ JSON ܇࿅ λΠϜϥΠϯ ϝσΟΞ σʔληοτ ܇࿅

Slide 20

Slide 20 text

Slide 21

Slide 21 text

إೝࣝ༻ͷϞσϧ DPO YY DPO YY GD QPPM Y DPO YY DPO YY QPPM Y GD PVUQVU MSO MSO Y

Slide 22

Slide 22 text

܇࿅ 12ສεςοϓֶशޙͷςετσʔλʹΑΔਖ਼౴཰ 96%

Slide 23

Slide 23 text

.FHBOF%FUFDUJPO Downloader Face Detection Megane Detection API TensorFlow ֬ೝɾमਖ਼ ೝࣝ݁Ռ JSON ܇࿅ λΠϜϥΠϯ ϝσΟΞ σʔληοτ ܇࿅

Slide 24

Slide 24 text

؟ڸೝࣝ༻ͷϞσϧ DPO YY DPO YY GD QPPM Y DPO YY DPO YY QPPM Y GD PVUQVU MSO MSO Y

Slide 25

Slide 25 text

܇࿅ 8ສεςοϓֶशޙͷςετσʔλʹΑΔਖ਼౴཰ 96%

Slide 26

Slide 26 text

C-LIS CO., LTD.

Slide 27

Slide 27 text

ೝࣝ݁Ռ Downloader Face Detection Megane Detection API TensorFlow ֬ೝɾमਖ਼ ೝࣝ݁Ռ JSON ܇࿅ λΠϜϥΠϯ ϝσΟΞ σʔληοτ ܇࿅

Slide 28

Slide 28 text

{ "generator": "Megane Co", "file_name": "neyuki_rei-CZY6tnUUYAAJVe6.jpg", "created_at": "2016-11-18T10:19:04.879281", "detected_faces": { "mode": "selective_search", "model_version": "20161118003546" "configurations": { }, "regions": [ { "rect": { "top": 59.0, "bottom": 175.0, "left": 85.0, "right": 176.0 }, "label": 2, "probability": 0.9962707757949829 }, ೝࣝ݁Ռͷ+40/ { "rect": { "top": 220.0, "bottom": 274.0, "left": 455.0, "right": 524.0 }, "label": 1, "probability": 0.9997575879096985 }, { "rect": { "top": 196.0, "bottom": 282.0, "left": 244.0, "right": 330.0 }, "label": 1, "probability": 1.0 }

Slide 29

Slide 29 text

ೝࣝ݁Ռͷ֬ೝͱमਖ਼ Downloader Face Detection Megane Detection API TensorFlow ֬ೝɾमਖ਼ ೝࣝ݁Ռ JSON ܇࿅ λΠϜϥΠϯ ϝσΟΞ σʔληοτ ܇࿅

Slide 30

Slide 30 text

C-LIS CO., LTD. https://github.com/keiji/region_cropper © ࠜઇΕ͍

Slide 31

Slide 31 text

ೝࣝ݁Ռͷ֬ೝͱमਖ਼ Downloader Face Detection Megane Detection API TensorFlow ֬ೝɾमਖ਼ ೝࣝ݁Ռ JSON ܇࿅ λΠϜϥΠϯ ϝσΟΞ σʔληοτ ܇࿅

Slide 32

Slide 32 text

ཷ·Δະॲཧը૾ ॲཧࡁΈɿ20,000ຕ ະॲཧɿ 24,000ຕ

Slide 33

Slide 33 text

ສ݅Λ௒͑ͨ͋ͨΓͰ  ํ਑Λมߋ

Slide 34

Slide 34 text

ೝࣝʹࣦഊͨ͠ը૾͚ͩ σʔληοτ΁௥Ճ

Slide 35

Slide 35 text

Ұ൪ࠨͷ؟ڸ່ͬͷإΛೝࣝͰ͖͍ͯͳ͍ © ࠜઇΕ͍

Slide 36

Slide 36 text

؟ڸΛ͔͚͍ͯͳ͍ͷʹ؟ڸ່ͬͱೝࣝ © ࠜઇΕ͍

Slide 37

Slide 37 text

إͱೝࣝ͢Δ৔ॴ͕ؒҧ͍ͬͯΔ ؟ڸ່ͬͱೝ͍ࣝͯ͠Δ © ࠜઇΕ͍

Slide 38

Slide 38 text

σʔληοτສຕʹ౸ୡ

Slide 39

Slide 39 text

C-LIS CO., LTD.

Slide 40

Slide 40 text

C-LIS CO., LTD.

Slide 41

Slide 41 text

C-LIS CO., LTD. ࢲͷ σʔληοτ͸ ສઍͰ͢ɻ

Slide 42

Slide 42 text

͗͢ΌʔΜϝϞ http://memo.sugyan.com/

Slide 43

Slide 43 text

C-LIS CO., LTD.

Slide 44

Slide 44 text

C-LIS CO., LTD. ࢲͷ σʔληοτ͸ ສઍͰ͢ɻ

Slide 45

Slide 45 text

΋͏ͪΐͬͱ͚ͩ ଓ͘Μ͡Ό

Slide 46

Slide 46 text

w؟ڸ່ͬը૾Λൃݟͨ͠Βϝϯγϣϯ͢ΔॲཧΛ࣮૷ wإೝࣝͷਫ਼౓޲্ ϞσϧͷܰྔԽ 4FMFDUJWF4FBSDIͷύϥϝʔλʔௐ੔ إͷύλʔϯʢਖ਼໘ɾԣإɾԼ޲͖ͳͲʣΛΫϥεʹ෼཭͢Δ wΠϥετإೝࣝͷ"1*Խ w؟ڸೝࣝͷΫϥεΛ૿΍͢ ࠓޙͷ՝୊

Slide 47

Slide 47 text

؟ڸೝࣝͷΫϥεΛ૿΍͢ʁ ԑͳ͠ ্ԑɾηϧϑϨʔϜ

Slide 48

Slide 48 text

ҧ͍·͢

Slide 49

Slide 49 text

޷Έͷʰ؟ڸ່ͬʱը૾Λ ࣗಈͰऩू͍ͨ͠

Slide 50

Slide 50 text

๻ʹʰ؟ڸ່ͬʱͷ࿩ͤͨ͞Β ௕͘ͳΓ·͢Αʁ

Slide 51

Slide 51 text

؟ڸೝࣝͷΫϥεΛ૿΍͢ 4PGUXBSF%FTJHO೥݄߸ʰ"OESPJE4UVEJPͷελΠϧͰޮ཰ΞοϓʱΑΓ © ࠜઇΕ͍

Slide 52

Slide 52 text

؟ڸೝࣝͷΫϥεΛ૿΍͢ © ࠜઇΕ͍ ϥϑஈ֊

Slide 53

Slide 53 text

؟ڸೝࣝͷΫϥεΛ૿΍͢ 4PGUXBSF%FTJHO೥݄߸ʰ"OESPJE4UVEJPͷελΠϧͰޮ཰ΞοϓʱΑΓ © ࠜઇΕ͍

Slide 54

Slide 54 text

Slide 55

Slide 55 text

ग़ྗ૚Λ௥Ճ͢Δ DPO YY DPO YY GD QPPM Y DPO YY DPO YY QPPM Y GD PVUQVU MSO MSO Y φΠε؟ڸ່ͬʂ ؟ڸΛ͔͚ͨঁࢠ Human

Slide 56

Slide 56 text

άϥϑͷΤΫεϙʔτ

Slide 57

Slide 57 text

5FOTPS'MPXͰߏஙͨ͠ʮάϥϑʯͱʮύϥϝʔλʔʯΛͦͷ··ॻ͖ग़͢ɻ γϦΞϥΠβʔʹ1SPUPDPM#V⒎FSTΛ࢖༻͍ͯ͠Δɻ άϥϑͷΤΫεϙʔτͱ͸ʁ

Slide 58

Slide 58 text

"OESPJEͳͲͰಡΈࠐΉ৔߹͸άϥϑͱύϥϝʔλʔΛΤΫεϙʔτ͢Δඞཁ ͕͋ΔʢΒ͍͠ʣɻ ࣥචͨ͠ͱ͖͸ɺ܇࿅݁ՌΛ֤ΤϙοΫ͝ͱʹΤΫεϙʔτͯ͠ɺ͋ͱͰҰؾ ʹਖ਼౴཰ΛݟΔͱ͍͏͜ͱΛͨ͠ɻ 4BWFSͱͷҧ͍

Slide 59

Slide 59 text

܇࿅ࡁΈͷϞσϧʢάϥϑʴύϥϝʔλʔʣ͕͋Δɻ άϥϑʹυϩοϓΞ΢τͳͲͷ܇࿅༻ͷॲཧؚ͕·Ε͍ͯΔɻ લఏ

Slide 60

Slide 60 text

def inference(image_node, keep_prob, batch_size=1):  conv1_1 = _conv('conv1_1', image_node, [3, 3, 3, 64], 64)  conv1_2 = _conv('conv1_2', conv1_1, [3, 3, 64, 64], 64)  pool1 = _pool('pool1', conv1_2, [1, 2, 2, 1])    conv2_1 = _conv('conv2_1', pool1, [3, 3, 64, 128], 128)  conv2_2 = _conv('conv2_2', conv2_1, [3, 3, 128, 128], 128)  pool2 = _pool('pool2', conv2_2, [1, 2, 2, 1])    conv3_1 = _conv('conv3_1', pool2, [3, 3, 128, 256], 256)  conv3_2 = _conv('conv3_2', conv3_1, [3, 3, 256, 256], 256)  pool3 = _pool('pool3', conv3_2, [1, 2, 2, 1])  ܇࿅༻ͷάϥϑ reshape = tf.reshape(pool3, [batch_size, -1])  dim = reshape.get_shape()[1].value    fc4 = _fc('fc4', reshape, [dim, 512], 512, keep_prob)  fc5 = _fc('fc5', fc4, [512, 256], 256, keep_prob)    # output  with tf.variable_scope('output') as scope:  weights = _get_weights(shape=[256, NUM_CLASSES], stddev=1 / 256.0)  biases = _get_biases([NUM_CLASSES], value=0.0)  logits = tf.add(tf.matmul(fc5, weights), biases, name='logits')    return logits 

Slide 61

Slide 61 text

def _fc(name, input_layer, weights_shape, biases_size, keep_prob):  with tf.variable_scope(name) as scope:  weights = _get_weights(shape=weights_shape, stddev=0.05)  biases = _get_biases([biases_size], value=0.1)  fc = tf.nn.relu(tf.matmul(input_layer, weights) + biases, name=scope.name)  fc = tf.nn.dropout(fc, keep_prob)  return fc  ܇࿅༻ͷάϥϑ http://goo.gl/forms/2IsX177eJ3CpNKng1 model_vgg_lite_minibatch.py

Slide 62

Slide 62 text

def _fc(name, input_layer, weights_shape, biases_size, keep_prob):  with tf.variable_scope(name) as scope:  weights = _get_weights(shape=weights_shape, stddev=0.05)  biases = _get_biases([biases_size], value=0.1)  fc = tf.nn.relu(tf.matmul(input_layer, weights) + biases, name=scope.name)  fc = tf.nn.dropout(fc, keep_prob)  return fc  ධՁ༻ͷάϥϑ http://goo.gl/forms/2IsX177eJ3CpNKng1 model_vgg_lite_minibatch_no_dropout.py

Slide 63

Slide 63 text

# coding: UTF-8  from __future__ import absolute_import  from __future__ import division  from __future__ import print_function    import tensorflow as tf    from tensorflow.python.framework import graph_util    import model_vgg_lite_minibatch_no_dropout as model    FLAGS = tf.app.flags.FLAGS  tf.app.flags.DEFINE_string('output_graph_path', None, "ग़ྗ͢ΔάϥϑͷσΟϨΫτϦ")  tf.app.flags.DEFINE_string('checkpoint_dir', './checkpoints', "νΣοΫϙΠϯτͷσΟϨΫτϦ")  tf.app.flags.DEFINE_integer('batch_size', 1, "όοναΠζ") 

Slide 64

Slide 64 text

def main(argv=None):  # eval༻ͷΤϯτϦϙΠϯτ  input_image = tf.placeholder(tf.float32, shape=[FLAGS.batch_size, 32, 32, 3], name='input_image')  logits = model.inference(input_image, batch_size=FLAGS.batch_size)    saver = tf.train.Saver()    with tf.Session() as sess:   # ύϥϝʔλʔͷಡΈࠐΈ  checkpoint = tf.train.get_checkpoint_state(FLAGS.checkpoint_dir)  if checkpoint and checkpoint.model_checkpoint_path:  saver.restore(sees, checkpoint.model_checkpoint_path)  # άϥϑʴύϥϝʔλʔͷΤΫεϙʔτ  constant_graph_def = graph_util.convert_variables_to_constants(  sess, sess.graph_def, ["output/logits"])  with tf.gfile.FastGFile(FLAGS.output_graph_path, 'wb') as f:  f.write(constant_graph_def.SerializeToString())

Slide 65

Slide 65 text

$ python3 convert_to_graph.py --checkpoint_dir ./checkpoints \ --output_graph_path ./graphs/exported_graph.pb \ --batch_size 1 Converted 18 variables to const ops. ධՁ༻άϥϑʴύϥϝʔλʔͷΤΫεϙʔτ

Slide 66

Slide 66 text

with tf.gfile.FastGFile(graph_file, 'rb') as f:  graph_def = tf.GraphDef()  graph_def.ParseFromString(f.read())  _ = tf.import_graph_def(graph_def, name='')  ධՁଆͰάϥϑΛΠϯϙʔτ

Slide 67

Slide 67 text

predictions = sess.run(logits,  feed_dict={  input_image: batch_byte_array,  labels: batch_label,  })  MPHJUTͱMBCFM͔Βਖ਼౴൑ఆ

Slide 68

Slide 68 text

labels = tf.placeholder(tf.int32, shape=[FLAGS.batch_size], name='label')  input_image = tf.get_default_graph().get_tensor_by_name('input_image:0')  logits = tf.get_default_graph().get_tensor_by_name('output/logits:0')    top_k_op = tf.nn.in_top_k(logits, labels, 1)  ΦϖϨʔγϣϯͷऔಘͱάϥϑͷߏங

Slide 69

Slide 69 text

࣮ߦ

Slide 70

Slide 70 text

E tensorflow/core/common_runtime/executor.cc:334] Executor failed to create kernel. Invalid argument: NodeDef mentions attr 'Tshape' not in Op output:T; attr=T:type>; NodeDef: Reshape = Reshape[T=DT_FLOAT, Tshape=DT_INT32, _device="/job:localhost/replica:0/task:0/cpu:0"] (pool3, Reshape/shape) [[Node: Reshape = Reshape[T=DT_FLOAT, Tshape=DT_INT32, _device="/job:localhost/ replica:0/task:0/cpu:0"](pool3, Reshape/shape)]] E tensorflow/core/client/tensor_c_api.cc:485] NodeDef mentions attr 'Tshape' not in Op output:T; attr=T:type>; NodeDef: Reshape = Reshape[T=DT_FLOAT, Tshape=DT_INT32, _device="/job:localhost/replica:0/task: 0/cpu:0"](pool3, Reshape/shape) [[Node: Reshape = Reshape[T=DT_FLOAT, Tshape=DT_INT32, _device="/job:localhost/ replica:0/task:0/cpu:0"](pool3, Reshape/shape)]] ಈ͖·ͤΜ

Slide 71

Slide 71 text

['input_image', 'conv1_1/weights', 'conv1_1/weights/read', 'conv1_1/Conv2D', 'conv1_1/biases', 'conv1_1/biases/read', 'conv1_1/BiasAdd', 'conv1_1/conv1_1', 'conv1_2/weights', 'conv1_2/weights/read', 'conv1_2/Conv2D', 'conv1_2/biases', 'conv1_2/biases/read', 'conv1_2/BiasAdd', 'conv1_2/conv1_2', 'pool1', ߏ੒͢Δϊʔυ 'conv2_1/weights', 'conv2_1/weights/read', 'conv2_1/Conv2D', 'conv2_1/biases', 'conv2_1/biases/read', 'conv2_1/BiasAdd', 'conv2_1/conv2_1', 'conv2_2/weights', 'conv2_2/weights/read', 'conv2_2/Conv2D', 'conv2_2/biases', 'conv2_2/biases/read', 'conv2_2/BiasAdd', 'conv2_2/conv2_2', 'pool2', 'conv3_1/weights', 'conv3_1/weights/read', 'conv3_1/Conv2D', 'conv3_1/biases', 'conv3_1/biases/read', 'conv3_1/BiasAdd', 'conv3_1/conv3_1', 'conv3_2/weights', 'conv3_2/weights/read', 'conv3_2/Conv2D', 'conv3_2/biases', 'conv3_2/biases/read', 'conv3_2/BiasAdd', 'conv3_2/conv3_2', 'pool3', 'Reshape/shape', 'Reshape', 'fc4/weights', 'fc4/weights/read', 'fc4/biases', 'fc4/biases/read', 'fc4/MatMul', 'fc4/add', 'fc4/fc4', 'fc5/weights', 'fc5/weights/read', 'fc5/biases', 'fc5/biases/read', 'fc5/MatMul', 'fc5/add', 'fc5/fc5', 'output/weights', 'output/weights/read', 'output/biases', 'output/biases/read', 'output/MatMul', 'output/logits']

Slide 72

Slide 72 text

['train_image', 'input_image', 'conv1_1/weights', 'conv1_1/weights/read', 'conv1_1/Conv2D', 'conv1_1/biases', 'conv1_1/biases/read', 'conv1_1/BiasAdd', 'conv1_1/conv1_1', 'conv1_2/weights', 'conv1_2/weights/read', 'conv1_2/Conv2D', 'conv1_2/biases', 'conv1_2/biases/read', 'conv1_2/BiasAdd', 'conv1_2/conv1_2', 'pool1', ܇࿅தʹΤΫεϙʔτͨ͠΋ͷͱൺֱ 'conv2_1/weights', 'conv2_1/weights/read', 'conv2_1/Conv2D', 'conv2_1/biases', 'conv2_1/biases/read', 'conv2_1/BiasAdd', 'conv2_1/conv2_1', 'conv2_2/weights', 'conv2_2/weights/read', 'conv2_2/Conv2D', 'conv2_2/biases', 'conv2_2/biases/read', 'conv2_2/BiasAdd', 'conv2_2/conv2_2', 'pool2', 'conv3_1/weights', 'conv3_1/weights/read', 'conv3_1/Conv2D', 'conv3_1/biases', 'conv3_1/biases/read', 'conv3_1/BiasAdd', 'conv3_1/conv3_1', 'conv3_2/weights', 'conv3_2/weights/read', 'conv3_2/Conv2D', 'conv3_2/biases', 'conv3_2/biases/read', 'conv3_2/BiasAdd', 'conv3_2/conv3_2', 'pool3', 'Reshape/shape', 'Reshape', 'fc4/weights', 'fc4/weights/read', 'fc4/biases', 'fc4/biases/read', 'fc4/MatMul', 'fc4/add', 'fc4/fc4', 'fc5/weights', 'fc5/weights/read', 'fc5/biases', 'fc5/biases/read', 'fc5/MatMul', 'fc5/add', 'fc5/fc5', 'output/weights', 'output/weights/read', 'output/biases', 'output/biases/read', 'output/MatMul', 'output/logits']

Slide 73

Slide 73 text

݁࿦ Α͘Θ͔Βͳ͍

Slide 74

Slide 74 text

/FYU ʰ5FOTPS'MPXΛ"OESPJEͰ࣮ߦ͠Α͏ʂʱ ݹ઒ɹ৽ ։ൃτϥοΫϧʔϜ5

Slide 75

Slide 75 text

ػցֶशϋϯζΦϯ https://gcpug-tokushima.connpass.com/event/43430/

Slide 76

Slide 76 text

C-LIS CO., LTD. ֤੡඼໊ɾϒϥϯυ໊ɺձ໊ࣾͳͲ͸ɺҰൠʹ֤ࣾͷ঎ඪ·ͨ͸ొ࿥঎ඪͰ͢ɻຊࢿྉதͰ͸ɺɺɺäΛׂѪ͍ͯ͠·͢ɻ ຊࢿྉ͸ɺ༗ݶձࣾγʔϦεͷஶ࡞෺Ͱ͢ɻܝࡌ͞Ε͍ͯΔΠϥετ͸ࠜઇΕ͍ͷஶ࡞෺Ͱ͢ɻ  ಺༰ͷશ෦ɺ·ͨ͸Ұ෦ʹ͍ͭͯɺஶ࡞ऀ͔ΒจॻʹΑΔڐ୚Λಘͣʹෳ੡͢Δ͜ͱ͸ې͡ΒΕ͍ͯ·͢ɻ