赞
踩
一、前言
我们有时候需要可视化特征图,尤其是发paper,或者对比算法等情况。而且通过可视化特征图,也可以让我们对这个整个cnn模型更加熟悉,废话不多说了。
二、效果图
下面我会给出代码,效果图分为单channel绘图和1:1通道特征图融合图。
我生成了很多特征图,我就简单的放两张吧,意思意思。
单通道特征图
叠加后的特征图
三、代码
我再次描述清楚我的需求以及我现有的东西,我有网络的结构和网络的预训练权重,我想通过输入图片,得到图片在网络特定层的特征图。
从main()开始看代码,我会说得详细一点,尽量让大家看懂, 这样你修改起来会方便很多。
图片保存和读取的路径相关的问题,我就不说了,这里大家应该懂。
1.首先我们看导入的包,DepthCompletionFrontNet 这是我的网络结构,首先你要搭建起的你的网络(这个得有)。
2.看main()函数,定位到get_feature()函数
3.get_feature做了下面得几个事儿,第一,读取图片,也就是要输入网络得图片(我得网络是双分支,所以是读取两个图,这里你读取一个图就行,就 img_rgb 就行,把 img_pc 相关内容注释);第二,定义网络,实例化,载入预训练权重模型;第三,定义我们要提取出得特定层,这里必须和你网络结构定义得层一模一样,一模一样,一模一样。
4.已经定义的网络结构需要进行修改,假设你网络定义的代码如下:
- # 仅仅举例子,我懒得补全了,直接csdn手打的
- class Net(nn.Module):
- super(Net,self).__init__()
- self.conv1 = nn.conv1
- self.conv2 = nn.conv2
- self.conv3 = nn.conv3
-
-
- forward(self,x):
- x = conv1(x)
- x = conv2(x)
- x = conv3(x)
- return x
网络的定义不需要修改,我们需要修改下网络的 forward,加入字典 all_dict去存储每层的tensor,forward修改如下:
- forward(self,x):
- all_dict = {}
-
- x = conv1(x)
- all_dict['conv1'] = x
- x = conv2(x)
- all_dict['conv2'] = x
- x = conv3(x)
- all_dict['conv3'] = x
- return x,all_dict
这样子就修改完成了
总结一下:首先读入模型和图片,图片在前向传播的过程中,我们通过字典保存每层的tensor,需要提取哪层,就从哪层去获取tensor,进而可视化。
大家有问题可以留言,我看到一定会回复。如可以运行,麻烦点赞下,谢谢!希望帮到大家。
完整代码如下(网络结构我的很复杂,就不放了, 网络结构修改就像上面我说的一样,你可以直接读取img_rgb,在模型的前向传播输入img_rgb,我的网络是双分支,所以我输入两个图组合的字典):
- import torch
- import torchvision.transforms as transforms
- import skimage.data
- import skimage.io
- import skimage.transform
- import numpy as np
- import matplotlib.pyplot as plt
- from completion_segmentation_model import DepthCompletionFrontNet
- # from completion_segmentation_model_v3_eca_attention import DepthCompletionFrontNet
- import math
-
- #https://blog.csdn.net/missyougoon/article/details/85645195
- # https://blog.csdn.net/grayondream/article/details/99090247
-
-
-
- # 定义是否使用GPU
- device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
-
-
-
- # 定义数据预处理方式(将输入的类似numpy中arrary形式的数据转化为pytorch中的张量(tensor))
- transform = transforms.ToTensor()
-
-
- def get_picture(picture_dir, transform):
- '''
- 该算法实现了读取图片,并将其类型转化为Tensor
- '''
- img = skimage.io.imread(picture_dir)
- img256 = skimage.transform.resize(img, (128, 256))
- img256 = np.asarray(img256)
- img256 = img256.astype(np.float32)
-
- return transform(img256)
-
-
- def get_picture_rgb(picture_dir):
- '''
- 该函数实现了显示图片的RGB三通道颜色
- '''
- img = skimage.io.imread(picture_dir)
- img256 = skimage.transform.resize(img, (256, 256))
- skimage.io.imsave('4.jpg', img256)
-
- # 取单一通道值显示
- # for i in range(3):
- # img = img256[:,:,i]
- # ax = plt.subplot(1, 3, i + 1)
- # ax.set_title('Feature {}'.format(i))
- # ax.axis('off')
- # plt.imshow(img)
-
- # r = img256.copy()
- # r[:,:,0:2]=0
- # ax = plt.subplot(1, 4, 1)
- # ax.set_title('B Channel')
- # # ax.axis('off')
- # plt.imshow(r)
-
- # g = img256.copy()
- # g[:,:,0]=0
- # g[:,:,2]=0
- # ax = plt.subplot(1, 4, 2)
- # ax.set_title('G Channel')
- # # ax.axis('off')
- # plt.imshow(g)
-
- # b = img256.copy()
- # b[:,:,1:3]=0
- # ax = plt.subplot(1, 4, 3)
- # ax.set_title('R Channel')
- # # ax.axis('off')
- # plt.imshow(b)
-
- # img = img256.copy()
- # ax = plt.subplot(1, 4, 4)
- # ax.set_title('image')
- # # ax.axis('off')
- # plt.imshow(img)
-
- img = img256.copy()
- ax = plt.subplot()
- ax.set_title('image')
- # ax.axis('off')
- plt.imshow(img)
-
- plt.show()
-
-
-
- def visualize_feature_map_sum(item,name):
- '''
- 将每张子图进行相加
- :param feature_batch:
- :return:
- '''
- feature_map = item.squeeze(0)
- c = item.shape[1]
- print(feature_map.shape)
- feature_map_combination=[]
- for i in range(0,c):
- feature_map_split = feature_map.data.cpu().numpy()[i, :, :]
-
- feature_map_combination.append(feature_map_split)
-
- feature_map_sum = sum(one for one in feature_map_combination)
- # feature_map = np.squeeze(feature_batch,axis=0)
- plt.figure()
- plt.title("combine figure")
- plt.imshow(feature_map_sum)
- plt.savefig('E:/Dataset/qhms/feature_map/feature_map_sum_'+name+'.png') # 保存图像到本地
- plt.show()
-
-
- def get_feature():
- # 输入数据
- root_path = 'E:/Dataset/qhms/data/small_data/'
- pic_dir = 'test_umm_000067.png'
- pc_path = root_path+'knn_pc_crop_0.6/'+pic_dir
- rgb_path = root_path+'train_image_2_lane_crop_0.6/'+pic_dir
-
-
- img_rgb = get_picture(rgb_path, transform)
- # 插入维度
- img_rgb = img_rgb.unsqueeze(0)
- img_rgb = img_rgb.to(device)
-
- img_pc = get_picture(pc_path, transform)
- # 插入维度
- img_pc = img_pc.unsqueeze(0)
- img_pc = img_pc.to(device)
-
-
- # 加载模型
- checkpoint = torch.load('E:/Dataset/qhms/all_result/v3/crop_0.6_old/hah/checkpoint-195.pth.tar')
- args = checkpoint['args']
- print(args)
- model = DepthCompletionFrontNet(args)
- print(model.keys())
- model.load_state_dict(checkpoint['model'])
- model.to(device)
-
-
-
- exact_list = ["conv1","conv2","conv3","conv4","convt4","convt3","convt2_","convt1_","lane"]
- # myexactor = FeatureExtractor(model, exact_list)
- img1 = {
- 'pc': img_pc, 'rgb': img_rgb
- }
- # print(img1['pc'])
- # x = myexactor(img1)
- result,all_dict = model(img1)
- outputs = []
-
-
- # 挑选exact_list的层
- for item in exact_list:
- x = all_dict[item]
- outputs.append(x)
-
-
- # 特征输出可视化
- x = outputs
- k=0
- print(x[0].shape[1])
- for item in x:
- c = item.shape[1]
-
- plt.figure()
- name = exact_list[k]
- plt.suptitle(name)
-
- for i in range(c):
- wid = math.ceil(math.sqrt(c))
- ax = plt.subplot(wid, wid, i + 1)
- ax.set_title('Feature {}'.format(i))
- ax.axis('off')
- figure_map = item.data.cpu().numpy()[0, i, :, :]
- plt.imshow(figure_map, cmap='jet')
- plt.savefig('E:/Dataset/qhms/feature_map/feature_map_' + name + '.png') # 保存图像到本地
- visualize_feature_map_sum(item,name)
- k = k + 1
- plt.show()
-
-
-
- # 训练
- if __name__ == "__main__":
- # get_picture_rgb(pic_dir)
- get_feature()
-
参考:
- https://blog.csdn.net/missyougoon/article/details/85645195
- https://blog.csdn.net/grayondream/article/details/99090247
Copyright © 2003-2013 www.wpsshop.cn 版权所有,并保留所有权利。