fix integrated system

Hironsan · Sep 14, 2016 · 11b1e14 · 11b1e14
1 parent 709a297
commit 11b1e14
Show file tree

Hide file tree

Showing 4 changed files with 67 additions and 85 deletions.
diff --git a/boss_input.py b/boss_input.py
@@ -4,8 +4,10 @@
 import numpy as np
 import cv2
 
+IMAGE_SIZE = 64
 
-def resize_with_pad(image, height, width):
+
+def resize_with_pad(image, height=IMAGE_SIZE, width=IMAGE_SIZE):
 
     def get_padding_size(image):
         h, w, _ = image.shape
@@ -24,7 +26,7 @@ def get_padding_size(image):
         return top, bottom, left, right
 
     top, bottom, left, right = get_padding_size(image)
-    BLACK = [0,0,0]
+    BLACK = [0, 0, 0]
     constant = cv2.copyMakeBorder(image, top , bottom, left, right, cv2.BORDER_CONSTANT, value=BLACK)
 
     resized_image = cv2.resize(constant, (height, width))
@@ -45,10 +47,8 @@ def traverse_dir(path):
                 image = read_image(abs_path)
                 images.append(image)
                 labels.append(path)
-    return images, labels
-
 
-IMAGE_SIZE = 32
+    return images, labels
 
 
 def read_image(file_path):
@@ -60,8 +60,8 @@ def read_image(file_path):
 
 def extract_data(path):
     images, labels = traverse_dir(path)
-    #images = np.array([np.reshape(image, -1) for image in images])
     images = np.array(images)
     dic = dict([(label, i) for i, label in enumerate(set(labels))])
     labels = np.array([dic[label] for label in labels])
+
     return images, labels
diff --git a/boss_train.py b/boss_train.py
@@ -3,7 +3,6 @@
 import random
 
 import numpy as np
-import tensorflow as tf
 from sklearn.cross_validation import train_test_split
 from keras.preprocessing.image import ImageDataGenerator
 from keras.models import Sequential
@@ -14,7 +13,7 @@
 from keras.models import load_model
 from keras import backend as K
 
-from boss_input import extract_data, resize_with_pad
+from boss_input import extract_data, resize_with_pad, IMAGE_SIZE
 
 
 class Dataset(object):
@@ -25,7 +24,7 @@ def __init__(self):
         self.Y_train = None
         self.Y_test = None
 
-    def read(self, img_rows=32, img_cols=32, img_channels=3, nb_classes=2):
+    def read(self, img_rows=IMAGE_SIZE, img_cols=IMAGE_SIZE, img_channels=3, nb_classes=2):
         images, labels = extract_data('./data/')
         labels = np.reshape(labels, [-1])
         # numpy.reshape
@@ -141,11 +140,12 @@ def load(self, file_path=FILE_PATH):
         self.model = load_model(file_path)
 
     def predict(self, image):
-        #result = self.model.predict_proba(image)
-        if image.shape != (1, 3, 32, 32):
-            image = resize_with_pad(image, 32, 32)
-            image = image.reshape((1, 3, 32, 32))
+        if image.shape != (1, 3, IMAGE_SIZE, IMAGE_SIZE):
+            image = resize_with_pad(image)
+            image = image.reshape((1, 3, IMAGE_SIZE, IMAGE_SIZE))
         result = self.model.predict_classes(image)
+        # result = self.model.predict_proba(image)
+
         return result[0]
 
     def evaluate(self, dataset):
@@ -154,21 +154,21 @@ def evaluate(self, dataset):
 
 if __name__ == '__main__':
 
-    #dataset = Dataset()
-    #dataset.read()
-    """
+    dataset = Dataset()
+    dataset.read()
+
     model = Model()
     model.build_model(dataset)
     model.train(dataset, nb_epoch=10)
     model.save()
-    """
+
     model = Model()
     model.load()
-    # model.evaluate(dataset)
-    # for image, label in zip(dataset.X_test, dataset.Y_test):
-    #     model.predict(image.reshape(1, 3, 32, 32))
-    #     print(label)
-
+    model.evaluate(dataset)
+    for image, label in zip(dataset.X_test, dataset.Y_test):
+        model.predict(image.reshape(1, 3, IMAGE_SIZE, IMAGE_SIZE))
+        print(label)
+    """
     import cv2
     import os
 
@@ -178,7 +178,8 @@ def evaluate(self, dataset):
         if file_name.endswith('.jpg'):
             print(file_name)
             image = cv2.imread('./data/boss/' + file_name)
-            image = resize_with_pad(image, 32, 32)
-            image = image.reshape((1, 3, 32, 32))
+            image = resize_with_pad(image)
+            image = image.reshape((1, 3, IMAGE_SIZE, IMAGE_SIZE))
             result = model.predict(image)
             print(result)
+    """
diff --git a/capture.py → camera_reader.py b/capture.py → camera_reader.py
@@ -1,6 +1,4 @@
 # -*- coding:utf-8 -*-
-#webカメラの映像から顔を探し白の枠線をつけて保存するプログラム
-
 import threading
 from datetime import datetime
 import cv2
@@ -68,46 +66,46 @@ def run(self):
 cv2.destroyAllWindows()
 """
 
+if __name__ == '__main__':
+    cap = cv2.VideoCapture(1)
+    cascade_path = "/usr/local/opt/opencv/share/OpenCV/haarcascades/haarcascade_frontalface_alt.xml"
+    model = Model()
+    model.load()
+    while True:
+        ret, frame = cap.read()
+        # グレースケール変換
+        frame_gray = cv2.cvtColor(frame, cv2.COLOR_BGR2GRAY)
 
-cap = cv2.VideoCapture(1)
-cascade_path = "/usr/local/opt/opencv/share/OpenCV/haarcascades/haarcascade_frontalface_alt.xml"
-model = Model()
-model.load()
-while True:
-    ret, frame = cap.read()
-    # グレースケール変換
-    frame_gray = cv2.cvtColor(frame, cv2.COLOR_BGR2GRAY)
-
-    # カスケード分類器の特徴量を取得する
-    cascade = cv2.CascadeClassifier(cascade_path)
-
-    # 物体認識（顔認識）の実行
-    facerect = cascade.detectMultiScale(frame_gray, scaleFactor=1.2, minNeighbors=3, minSize=(10, 10))
-    if len(facerect) > 0:
-        print('face detected')
-        color = (255, 255, 255)  # 白
-        for rect in facerect:
-            # 検出した顔を囲む矩形の作成
-            #cv2.rectangle(frame, tuple(rect[0:2]), tuple(rect[0:2] + rect[2:4]), color, thickness=2)
-
-            x, y = rect[0:2]
-            width, height = rect[2:4]
-            image = frame[y-50: y + height, x: x + width + 50]
-            cv2.imwrite('test.jpg', image)
-            result = model.predict(image)
-            print(result)
-            if result == 1:  # boss
-                print('Boss is approaching')
-                show_image()
-            else:
-                print('Not boss')
-
-    #10msecキー入力待ち
-    k = cv2.waitKey(100)
-    #Escキーを押されたら終了
-    if k == 27:
-        break
+        # カスケード分類器の特徴量を取得する
+        cascade = cv2.CascadeClassifier(cascade_path)
 
-#キャプチャを終了
-cap.release()
-cv2.destroyAllWindows()
+        # 物体認識（顔認識）の実行
+        facerect = cascade.detectMultiScale(frame_gray, scaleFactor=1.2, minNeighbors=3, minSize=(10, 10))
+        if len(facerect) > 0:
+            print('face detected')
+            color = (255, 255, 255)  # 白
+            for rect in facerect:
+                # 検出した顔を囲む矩形の作成
+                #cv2.rectangle(frame, tuple(rect[0:2]), tuple(rect[0:2] + rect[2:4]), color, thickness=2)
+
+                x, y = rect[0:2]
+                width, height = rect[2:4]
+                image = frame[y-50: y + height, x: x + width + 50]
+                cv2.imwrite('test.jpg', image)
+                result = model.predict(image)
+                print(result)
+                if result == 1:  # boss
+                    print('Boss is approaching')
+                    show_image()
+                else:
+                    print('Not boss')
+
+        #10msecキー入力待ち
+        k = cv2.waitKey(100)
+        #Escキーを押されたら終了
+        if k == 27:
+            break
+
+    #キャプチャを終了
+    cap.release()
+    cv2.destroyAllWindows()
diff --git a/image_show.py b/image_show.py
@@ -1,27 +1,10 @@
 # -*- coding: utf-8 -*-
 import sys
 
-import cv2
 from PyQt4 import QtGui
 
 
 def show_image(image_path='s_pycharm.jpg'):
-    """
-    # 画像の読み込み
-    RGB = 1
-    img = cv2.imread(image_path, RGB)
-
-    # 画像の表示
-    cv2.imshow('img', img)
-
-    # キーが押させるまで画像を表示したままにする
-    # 第一引数：キーイベントを待つ時間　0: 無限, 0以上: 指定ミリ秒待つ
-    cv2.waitKey(0)
-
-    # 作成したウィンドウを全て破棄
-    cv2.destroyAllWindows()
-    """
-
     app = QtGui.QApplication(sys.argv)
     pixmap = QtGui.QPixmap(image_path)
     screen = QtGui.QLabel()
@@ -31,4 +14,4 @@ def show_image(image_path='s_pycharm.jpg'):
 
 
 if __name__ == '__main__':
-    show_image()
+    show_image()