lfw 数据集上的 vgg16 不获取 RGB

Posted 2023-03-16

【中文标题】lfw 数据集上的 vgg16 不获取 RGB

【英文标题】：vgg16 on lfw dataset Dont get RGB

【发布时间】：2020-10-08 17:37:05

【问题描述】：

我想在 LFW 数据集上使用 Keras 库中的 VGGNET-16 进行人脸识别。但是生成的是灰度而不是 RGB，我无法更改它：

【问题讨论】：

嗨@MahboobehNajafi，虽然我已经为您展示了图像，但它通常不是很好，因为我们无法重现您遇到的错误。您能否输入您使用的代码，如果可能，提供一个可重现的示例，以便我们查看您的错误来自何处并进行故障排除？

【参考方案1】：

      #load dataset
 lfw_people = fetch_lfw_people(min_faces_per_person=53, resize=0.4)


 

# introspect the images arrays to find the shapes (for plotting)
n_samples, h, w = lfw_people.images.shape

# for machine learning we use the 2 data directly (as relative pixel
# positions info is ignored by this model)
# access the images

X = lfw_people.data
n_features = X.shape[1]


# the label to predict is the id of the person
# access the class labels
y = lfw_people.target
target_names = lfw_people.target_names
n_classes = target_names.shape[0]



print("Total dataset size:")
print("number of samples: " , n_samples)
print("number of classes: " , n_classes)
print("image dimensions: ", h, w)
print("number of features per image: ", h*w)


lfw_people.target_names

** 数组（['阿里尔沙龙'，'科林鲍威尔'，'唐纳德拉姆斯菲尔德'，'乔治W布什'， “格哈德·施罗德”、“雨果·查韦斯”、“让·克雷蒂安”、 '约翰阿什克罗夫特'，'小泉纯一郎'，'托尼布莱尔']，dtype ='

  labelNames=["Ariel Sharon", "Colin Powell", "Donald Rumsfeld", "George W Bush",
           "Gerhard Schroeder", "Hugo Chavez", "Jean Chretien",
       Ahcroft", "Junichiro Koizumi", "Tony Blair"]
    X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.4)
print(X_train.shape)

**

(873, 1850)

**

   y_train= to_categorical(y_train, num_classes= 10)
    y_test = to_categorical(y_test,num_classes= 10)
    chanDim=1

改变大小

    img=[]
    for i in range(len(X_train)):
      z=cv2.resize(X_train[i],(224,224)).astype(np.float32)
      z = np.expand_dims(z, axis=0)
      img.append(z)
Xnew_train=np.array(img)
img=[]
for i in range(len(X_test)):
  r=cv2.resize(X_train[i],(224,224)).astype(np.float32)
  r = np.expand_dims(r, axis=0)
  img.append(r)
 Xnew_test=np.array(img)
     

这里我得到 1 的 3。bc LFW 是 RGB Xnew_train.shape

**

(873, 1, 224, 224)

**

    model = Sequential()
model.add(ZeroPadding2D((1,1),input_shape=(3,224,224)))
model.add(Convolution2D(64, 3, 3, activation='relu'))
model.add(BatchNormalization(axis=chanDim))
model.add(ZeroPadding2D((1,1)))
model.add(Convolution2D(64, 3, 3, activation='relu'))
model.add(BatchNormalization(axis=chanDim))
model.add(MaxPooling2D(pool_size=(2, 2), dim_ordering="th" , strides=(2,2)))

model.add(ZeroPadding2D((1,1)))
model.add(Convolution2D(128, 3, 3, activation='relu'))
model.add(BatchNormalization(axis=chanDim))
model.add(ZeroPadding2D((1,1)))
model.add(Convolution2D(128, 3, 3, activation='relu'))
model.add(BatchNormalization(axis=chanDim))
model.add(MaxPooling2D(pool_size=(2, 2), dim_ordering="th" , strides=(2,2)))

model.add(ZeroPadding2D((1,1)))
model.add(Convolution2D(256, 3, 3, activation='relu'))
model.add(BatchNormalization(axis=chanDim))
model.add(ZeroPadding2D((1,1)))
model.add(Convolution2D(256, 3, 3, activation='relu'))
model.add(BatchNormalization(axis=chanDim))
model.add(ZeroPadding2D((1,1)))
model.add(Convolution2D(256, 3, 3, activation='relu'))
model.add(BatchNormalization(axis=chanDim))
model.add(MaxPooling2D(pool_size=(2, 2), dim_ordering="th" , strides=(2,2)))

model.add(ZeroPadding2D((1,1)))
model.add(Convolution2D(256, 3, 3, activation='relu'))
model.add(BatchNormalization(axis=chanDim))
model.add(ZeroPadding2D((1,1)))
model.add(Convolution2D(256, 3, 3, activation='relu'))
model.add(BatchNormalization(axis=chanDim))
model.add(ZeroPadding2D((1,1)))
model.add(Convolution2D(256, 3, 3, activation='relu'))
model.add(BatchNormalization(axis=chanDim))
model.add(MaxPooling2D(pool_size=(2, 2), dim_ordering="th" , strides=(2,2)))

model.add(ZeroPadding2D((1,1)))
model.add(Convolution2D(256, 3, 3, activation='relu'))
model.add(ZeroPadding2D((1,1)))
model.add(Convolution2D(256, 3, 3, activation='relu'))
model.add(ZeroPadding2D((1,1)))
model.add(Convolution2D(256, 3, 3, activation='relu'))
model.add(MaxPooling2D(pool_size=(2, 2), dim_ordering="th" , strides=(2,2)))

model.add(Flatten())
model.add(Dense(512, activation='relu', kernel_regularizer=regularizers.l2(0.01)))
model.add(Dropout(0.5))
model.add(Dense(512, activation='relu', kernel_regularizer=regularizers.l2(0.01)))
model.add(Dropout(0.5))
model.add(Dense(10, activation='softmax'))


 model.summary()
# Let's train the model using RMSprop
model.compile(loss='categorical_crossentropy',
         



 optimizer="Adam",
              metrics=['accuracy'])
#model.summary()

这里出现错误，如果在 input_shape=(3,224,224) 中使用 3 中的 1，则测试准确度 很低

    #NUM_EPOCHS = 25
BS = 32
batch_size= 128
history=model.fit(Xnew_train,y_train,batch_size=batch_size,epochs=200,validation_split=0.2,verbose=1)

**错误是：

**

ValueError: Error when checking input: expected zero_padding2d_1_input to have shape (3, 224, 224) but got array with shape (1, 224, 224)