fix bug

.gitignore

@@ -5,3 +5,5 @@ cv/__pycache__/
 cv/static/FaceUpload/
 dm/__pycache__/
 nlp/__pycache__/
+__pycache__/
+pyWeb/

README.md

@@ -34,7 +34,7 @@ with RESTful APIs. The platform is developed and maintained by [@LucasX](https:/
     - [x] Age Estimation 
 
 ## Deployment
-1. create a virtual enviroment named ```pyWeb```
+1. create a virtual enviroment named ```pyWeb``` follow [this tutorial](https://www.liaoxuefeng.com/wiki/0014316089557264a6b348958f449949df42a6d3a2e542c000/001432712108300322c61f256c74803b43bfd65c6f8d0d0000)
 2. install [Django](https://docs.djangoproject.com/en/2.1/intro/install/) and [PyTorch](https://pytorch.org/)
 3. install all dependent libraries: ```pip3 install -r requirements.txt```
 4. activate Python Web environment: ```source ~/pyWeb/bin/activate pyWeb```

XCloud/settings.py

@@ -119,4 +119,5 @@
 
 STATIC_ROOT = ''
 STATIC_URL = '/static/'
-STATICFILES_DIRS = ('D:/Users/LucasX/PycharmProjects/XCloud/cv/static', '/root/Project/XCloud/cv/static')
+STATICFILES_DIRS = ('D:/Users/LucasX/PycharmProjects/XCloud/cv/static', '/root/Project/XCloud/cv/static',
+                    'D:/Users/29140\PycharmProjects\XCloud\cv\static')

cv/controllers.py

@@ -1,18 +1,21 @@
 import json
 import os
 import time
-from random import randint
 
+import numpy as np
 import cv2
 from django.http import HttpResponse
 from mtcnn.mtcnn import MTCNN
+from sklearn.externals import joblib
+
+from cv import features
 
 BASE_DIR = os.path.dirname(os.path.dirname(os.path.abspath(__file__)))
 
 URL_PORT = 'http://localhost:8000'
 
 
-def upload_and_rec(request):
+def upload_and_rec_beauty(request):
     """
     upload and recognize image
     :param request:
@@ -43,11 +46,13 @@ def upload_and_rec(request):
             tik = time.time()
             imagepath = URL_PORT + '/static/FaceUpload/' + image.name
 
+            beauty = BeautyRecognizer().infer(os.path.join(image_dir, image.name))
+
             result['code'] = 0
             result['msg'] = 'success'
             result['data'] = {
                 'imgpath': imagepath,
-                'beauty': randint(0, 9)
+                'beauty': beauty
             }
             result['elapse'] = time.time() - tik
 
@@ -78,14 +83,27 @@ def detect_face(img_path):
 
 
 class BeautyRecognizer:
-    def __init__(self, pretrained_model):
-        self.model = None
+    def __init__(self, pretrained_model='./model/GradientBoostingRegressor.pkl'):
+        gbr = joblib.load(pretrained_model)
+        self.model = gbr
 
     def infer(self, img_path):
         img = cv2.imread(img_path)
         mtcnn_result = detect_face(img_path)
-        bbox = mtcnn_result['box']
-        face_region = img[bbox[0] - int(bbox[2] / 2): bbox[0] + int(bbox[2] / 2),
-                      bbox[1] - int(bbox[3] / 2): bbox[1] + int(bbox[3] / 2)]
+        bbox = mtcnn_result[0]['box']
+
+        margin_pixel = 10
+        face_region = img[bbox[0] - margin_pixel: bbox[0] + bbox[2] + margin_pixel,
+                      bbox[1] - margin_pixel: bbox[1] + bbox[3] + margin_pixel]
+
+        ratio = max(face_region.shape[0], face_region.shape[1]) / min(face_region.shape[0], face_region.shape[1])
+        if face_region.shape[0] < face_region.shape[1]:
+            face_region = cv2.resize(face_region, (int(ratio * 64), 64))
+            face_region = face_region[:,
+                          int((face_region.shape[0] - 64) / 2): int((face_region.shape[0] - 64) / 2) + 64]
+        else:
+            face_region = cv2.resize(face_region, (64, int(ratio * 64)))
+            face_region = face_region[int((face_region.shape[1] - 64) / 2): int((face_region.shape[1] - 64) / 2) + 64,
+                          :]
 
-        return self.model.infer(face_region)
+        return self.model.predict(np.array(features.HOG_from_cv(face_region).reshape(1, -1)))[0]

cv/features.py

@@ -82,6 +82,13 @@ def RAW(img_path):
     return img.reshape(img.shape[0] * img.shape[1])
 
 
+def RAW_from_cv(img):
+    img = skimage.color.rgb2gray(img)
+    img = (img - np.mean(img)) / np.std(img)
+
+    return img.reshape(img.shape[0] * img.shape[1])
+
+
 def HOG_from_cv(img):
     """
     extract HOG feature from opencv image object

cv/templates/index.html

@@ -49,7 +49,7 @@ <h3>From Research to Production</h3>
         <div class="row">
             <div class="col-lg-12 text-center">
                 <h2>HMT-Net</h2>
-                <p class="lead">Analyse Your Face in <span style="color: red;
+                <p class="lead">Analyze Your Face in <span style="color: red;
                 font-size:larger">Real-time</span></p>
             </div>
         </div>

cv/train_and_test_fbp_with_ml.py

@@ -6,16 +6,17 @@
 
 import pandas as pd
 from mtcnn.mtcnn import MTCNN
+from sklearn.ensemble import GradientBoostingRegressor
 from sklearn.externals import joblib
-from sklearn.model_selection import train_test_split
-from sklearn.ensemble import RandomForestRegressor
 from sklearn.metrics import mean_absolute_error, mean_squared_error
+from sklearn.model_selection import train_test_split
 
 sys.path.append('../')
 from cv.features import *
 
-SCUT5500_DIR = "E:\DataSet\CV\SCUT-FBP5500\Images"
-LABEL_CSV = "E:\DataSet\CV\SCUT-FBP5500/train_test_files\SCUT-FBP5500-With-Head.csv"
+SCUT5500_DIR = "E:\DataSet\Face\SCUT-FBP5500\Images"
+SCUT5500_CROP_DIR = "E:\DataSet\Face\SCUT-FBP5500\Crop"
+LABEL_CSV = "E:\DataSet\Face\SCUT-FBP5500/train_test_files\SCUT-FBP5500-With-Head.csv"
 
 
 def prepare_data():
@@ -54,8 +55,8 @@ def prepare_data():
             face_region = cv2.resize(img, (64, 64))
 
         lbp = LBP_from_cv(face_region)
-        hog = HOG_from_cv(face_region)
-        ldmk = Geo_from_cv(img)
+        # hog = HOG_from_cv(face_region)
+        # ldmk = Geo_from_cv(img)
 
         feature = lbp
         features.append(feature)
@@ -64,21 +65,50 @@ def prepare_data():
     return features, lbs
 
 
+def prepare_data_from_crop():
+    """
+    prepare face datasets from cropped facial regions
+    :return:
+    """
+    features = []
+    lbs = []
+
+    df = pd.read_csv(LABEL_CSV)
+    files = df['file']
+    scores = df['score']
+
+    for i in range(len(files)):
+        img = cv2.imread(os.path.join(SCUT5500_CROP_DIR, files[i]))
+
+        face_region = cv2.resize(img, (64, 64))
+
+        hog = HOG_from_cv(face_region)
+        # raw = RAW_from_cv(face_region)
+        # lbp = LBP_from_cv(face_region)
+        # ldmk = Geo_from_cv(img)
+
+        feature = list(hog)
+        features.append(feature)
+        lbs.append(scores[i])
+
+    return features, lbs
+
+
 def train_fbp(X, y):
     X_train, X_test, y_train, y_test = train_test_split(
         X, y, test_size=0.2, random_state=42)
 
-    rfreg = RandomForestRegressor()
-    print('start training Random Forest Regressor...')
-    rfreg.fit(X_train, y_train)
+    gbr = GradientBoostingRegressor()
+    print('start training regressor...')
+    gbr.fit(X_train, y_train)
 
     if not os.path.exists('./model'):
         os.makedirs('./model')
 
-    joblib.dump(rfreg, './model/RandomForestRegressor.pkl')
+    joblib.dump(gbr, './model/GradientBoostingRegressor.pkl')
     print('The regression model has been persisted...')
 
-    y_pred = rfreg.predict(X_test)
+    y_pred = gbr.predict(X_test)
 
     mae_lr = round(mean_absolute_error(y_test, np.array(y_pred).ravel()), 4)
     rmse_lr = round(np.math.sqrt(mean_squared_error(np.array(y_test), np.array(y_pred).ravel())), 4)
@@ -90,5 +120,5 @@ def train_fbp(X, y):
 
 
 if __name__ == '__main__':
-    X, y = prepare_data()
+    X, y = prepare_data_from_crop()
     train_fbp(X, y)

cv/views.py

@@ -28,7 +28,7 @@ def index(request):
 
 def fbp(request):
     from cv import controllers
-    return controllers.upload_and_rec(request)
+    return controllers.upload_and_rec_beauty(request)
 
 
 def fbp_view(request):

requirements.txt

@@ -1,5 +1,5 @@
 jieba==0.39
 opencv_python==3.4.0.12
 mtcnn==0.0.6
-Django==2.1.3
+Django==2.1.4
 snownlp==0.12.3