有哪些做封面的网站,代理贷款平台加盟,网站开发完整教程,自建网站编程学习参考来自
CNN可视化Convolutional Featureshttps://github.com/wmn7/ML_Practice/blob/master/2019_05_27/filter_visualizer.ipynb 文章目录 filter 的激活值 filter 的激活值
原理#xff1a;找一张图片#xff0c;使得某个 layer 的 filter 的激活值最大#xff0c…
学习参考来自
CNN可视化Convolutional Featureshttps://github.com/wmn7/ML_Practice/blob/master/2019_05_27/filter_visualizer.ipynb 文章目录 filter 的激活值 filter 的激活值
原理找一张图片使得某个 layer 的 filter 的激活值最大这张图片就是能被这个 filter 所检测的对象。
来个案例流程
初始化一张图片, 56X56使用预训练好的 VGG16 网络固定网络参数若想可视化第 40 层 layer 的第 k 个 filter 的 conv, 我们设置 loss 函数为 (-1*神经元激活值)梯度下降, 对初始图片进行更新对得到的图片X1.2, 得到新的图片重复上面的步骤
其中第五步比较关键我们可以看到初始化的图片不是很大只有56X56. 这是因为原文作者在实际做的时候发现若初始图片较大得到的特征的频率会较高即没有现在这么好的显示效果。
import torch
from torch.autograd import Variable
from PIL import Image, ImageOps
import torchvision.transforms as transforms
import torchvision.models as modelsimport numpy as np
import cv2
from cv2 import resize
from matplotlib import pyplot as pltdevice torch.device(cuda if torch.cuda.is_available() else cpu)initialize input image
sz 56
img np.uint(np.random.uniform(150, 180, (3, sz, sz))) / 255 # (3, 56, 56)
img torch.from_numpy(img[None]).float().to(device) # (1, 3, 56, 56)pretrained model
model_vgg16 models.vgg16_bn(pretrainedTrue).features.to(device).eval()
# downloading /home/xxx/.cache/torch/hub/checkpoints/vgg16_bn-6c64b313.pth, 500M
# print(model_vgg16)
# print(len(list(model_vgg16.children()))) # 44
# print(list(model_vgg16.children()))get the filters output of one layer
# 使用hook来得到网络中间层的输出
class SaveFeatures():def __init__(self, module):self.hook module.register_forward_hook(self.hook_fn)def hook_fn(self, module, input, output):self.features output.clone()def close(self):self.hook.remove()layer 42
activations SaveFeatures(list(model_vgg16.children())[layer])backpropagation, setting hyper-parameters
lr 0.1
opt_steps 25 # 迭代次数
filters 265 # layer 42 的第 265 个 filter使其激活值最大
upscaling_steps 13 # 图像放大次数
blur 3
upscaling_factor 1.2 # 放大倍率preprocessing of datasets
cnn_normalization_mean torch.tensor([0.485, 0.456, 0.406]).view(-1, 1, 1).to(device)
cnn_normalization_std torch.tensor([0.299, 0.224, 0.225]).view(-1, 1, 1).to(device)gradient descent
for epoch in range(upscaling_steps): # scale the image up up_scaling_steps timesimg (img - cnn_normalization_mean) / cnn_normalization_stdimg[img 1] 1img[img 0] 0print(Image Shape1:, img.shape)img_var Variable(img, requires_gradTrue) # convert image to Variable that requires gradoptimizeroptimizer torch.optim.Adam([img_var], lrlr, weight_decay1e-6)for n in range(opt_steps):optimizer.zero_grad()model_vgg16(img_var) # forwardloss -activations.features[0, filters].mean() # max the activationsloss.backward()optimizer.step()restore the imageprint(Loss:, loss.cpu().detach().numpy())img img_var * cnn_normalization_std cnn_normalization_meanimg[img1] 1img[img0] 0img img.data.cpu().numpy()[0].transpose(1,2,0)sz int(upscaling_factor * sz) # calculate new image sizeimg cv2.resize(img, (sz, sz), interpolationcv2.INTER_CUBIC) # scale image upif blur is not None:img cv2.blur(img, (blur, blur)) # blur image to reduce high frequency patternsprint(Image Shape2:, img.shape)img torch.from_numpy(img.transpose(2, 0, 1)[None]).to(device)print(Image Shape3:, img.shape)print(str(epoch), , Finished)print(*10)activations.close() # remove the hookimage img.cpu().clone()
image image.squeeze(0)
unloader transforms.ToPILImage()image unloader(image)
image cv2.cvtColor(np.asarray(image), cv2.COLOR_RGB2BGR)
cv2.imwrite(res1.jpg, image)
torch.cuda.empty_cache()
Image Shape1: torch.Size([1, 3, 56, 56])
Loss: -6.0634975
Image Shape2: (67, 67, 3)
Image Shape3: torch.Size([1, 3, 67, 67])
0 , FinishedImage Shape1: torch.Size([1, 3, 67, 67])
Loss: -7.8898916
Image Shape2: (80, 80, 3)
Image Shape3: torch.Size([1, 3, 80, 80])
1 , FinishedImage Shape1: torch.Size([1, 3, 80, 80])
Loss: -8.730318
Image Shape2: (96, 96, 3)
Image Shape3: torch.Size([1, 3, 96, 96])
2 , FinishedImage Shape1: torch.Size([1, 3, 96, 96])
Loss: -9.697872
Image Shape2: (115, 115, 3)
Image Shape3: torch.Size([1, 3, 115, 115])
3 , FinishedImage Shape1: torch.Size([1, 3, 115, 115])
Loss: -10.190881
Image Shape2: (138, 138, 3)
Image Shape3: torch.Size([1, 3, 138, 138])
4 , FinishedImage Shape1: torch.Size([1, 3, 138, 138])
Loss: -10.315895
Image Shape2: (165, 165, 3)
Image Shape3: torch.Size([1, 3, 165, 165])
5 , FinishedImage Shape1: torch.Size([1, 3, 165, 165])
Loss: -9.73861
Image Shape2: (198, 198, 3)
Image Shape3: torch.Size([1, 3, 198, 198])
6 , FinishedImage Shape1: torch.Size([1, 3, 198, 198])
Loss: -9.503629
Image Shape2: (237, 237, 3)
Image Shape3: torch.Size([1, 3, 237, 237])
7 , FinishedImage Shape1: torch.Size([1, 3, 237, 237])
Loss: -9.488493
Image Shape2: (284, 284, 3)
Image Shape3: torch.Size([1, 3, 284, 284])
8 , FinishedImage Shape1: torch.Size([1, 3, 284, 284])
Loss: -9.100454
Image Shape2: (340, 340, 3)
Image Shape3: torch.Size([1, 3, 340, 340])
9 , FinishedImage Shape1: torch.Size([1, 3, 340, 340])
Loss: -8.699549
Image Shape2: (408, 408, 3)
Image Shape3: torch.Size([1, 3, 408, 408])
10 , FinishedImage Shape1: torch.Size([1, 3, 408, 408])
Loss: -8.90135
Image Shape2: (489, 489, 3)
Image Shape3: torch.Size([1, 3, 489, 489])
11 , FinishedImage Shape1: torch.Size([1, 3, 489, 489])
Loss: -8.838546
Image Shape2: (586, 586, 3)
Image Shape3: torch.Size([1, 3, 586, 586])
12 , Finished
Process finished with exit code 0得到特征图 网上找个图片测试下看响应是不是最大
测试图片 import torch
from torch.autograd import Variable
from PIL import Image, ImageOps
import torchvision.transforms as transforms
import torchvision.models as modelsimport numpy as np
import cv2
from cv2 import resize
from matplotlib import pyplot as pltdevice torch.device(cuda if torch.cuda.is_available() else cpu)class SaveFeatures():def __init__(self, module):self.hook module.register_forward_hook(self.hook_fn)def hook_fn(self, module, input, output):self.features output.clone()def close(self):self.hook.remove()size (224, 224)
picture Image.open(./bird.jpg).convert(RGB)
picture ImageOps.fit(picture, size, Image.ANTIALIAS)loader transforms.ToTensor()
picture loader(picture).to(device)
print(picture.shape)cnn_normalization_mean torch.tensor([0.485, 0.456, 0.406]).view(-1, 1, 1).to(device)
cnn_normalization_std torch.tensor([0.229, 0.224, 0.225]).view(-1, 1, 1).to(device)picture (picture-cnn_normalization_mean) / cnn_normalization_stdmodel_vgg16 models.vgg16_bn(pretrainedTrue).features.to(device).eval()
print(list(model_vgg16.children())[40]) # Conv2d(512, 512, kernel_size(3, 3), stride(1, 1), padding(1, 1))
print(list(model_vgg16.children())[41]) # BatchNorm2d(512, eps1e-05, momentum0.1, affineTrue, track_running_statsTrue)
print(list(model_vgg16.children())[42]) # ReLU(inplaceTrue)layer 42
filters 265
activations SaveFeatures(list(model_vgg16.children())[layer])with torch.no_grad():picture_var Variable(picture[None])model_vgg16(picture_var)
activations.close()print(activations.features.shape) # torch.Size([1, 512, 14, 14])# 画出每个 filter 的平均值
mean_act [activations.features[0, i].mean().item() for i in range(activations.features.shape[1])]
plt.figure(figsize(7,5))
act plt.plot(mean_act, linewidth2.)
extraticks [filters]
ax act[0].axes
ax.set_xlim(0, 500)
plt.axvline(xfilters, colorgray, linestyle--)
ax.set_xlabel(feature map)
ax.set_ylabel(mane activation)
ax.set_xticks([0, 200, 400] extraticks)
plt.show()
torch.Size([3, 224, 224])
Conv2d(512, 512, kernel_size(3, 3), stride(1, 1), padding(1, 1))
BatchNorm2d(512, eps1e-05, momentum0.1, affineTrue, track_running_statsTrue)
ReLU(inplaceTrue)
torch.Size([1, 512, 14, 14])可以看到265 特征图对该输入的相应最高
总结实测了其他 layer 和 filter画出来的直方图中对应的 filter 相应未必是最高的不过也很高可能找的待测图片并不是最贴合设定 layer 的某个 filter 的特征。
