blocks|key|84417|text|VGG16模型本身只是固定的层序列和固定的卷积核大小等的一组权重，这并不意味着这些卷积核不能应用于其他大小的图像。|type|unstyled|depth|inlineStyleRanges|offset|length|style|CODE|entityRanges|data|84418|例如，在您的案例中：|84419|from+keras.models+import+Model
from+keras.layers+import+Dense,Flatten
from+keras.applications+import+vgg16
from+keras+import+backend+as+K

model+=+vgg16.VGG16(weights='imagenet',+include_top=False,+input_shape=(160,320,3))
model.summary(line_length=150)

flatten+=+Flatten()
new_layer2+=+Dense(10,+activation='softmax',+name='my_dense_2')

inp2+=+model.input
out2+=+new_layer2(flatten(model.output))

model2+=+Model(inp2,+out2)
model2.summary(line_length=150)|code-block|syntax|javascript|84420|根据here的说法，最小图像大小可以是48x48x3，超过这个值的任何值都可以。|84421|现在，原始权重确实是在224,224,3形状的图像上学习的，但过滤器权重对于新图像集的新任务来说是非常好的起点。您确实需要重新训练网络，但网络会非常快地收敛。这是迁移学习的基础。|84422|entityMap|0|LINK|mutability|MUTABLE|url|https://github.com/fchollet/keras/blob/master/keras/applications/vgg16.py#L77^0|0|5|0|0|0|J|7|2|4|0|0|B|9|0^^$0|@$1|2|3|4|5|6|7|Y|8|@$9|Z|A|10|B|C]]|D|@]|E|$]]|$1|F|3|G|5|6|7|11|8|@]|D|@]|E|$]]|$1|H|3|I|5|J|7|12|8|@]|D|@]|E|$K|L]]|$1|M|3|N|5|6|7|13|8|@$9|14|A|15|B|C]]|D|@$9|16|A|17|1|18]]|E|$]]|$1|O|3|P|5|6|7|19|8|@$9|1A|A|1B|B|C]]|D|@]|E|$]]|$1|Q|3|-4|5|6|7|1C|8|@]|D|@]|E|$]]]|R|$S|$5|T|U|V|E|$W|X]]]]

<code>VGG16</code> model in itself is just a set of weights of the fixed sequence of layers and fixed convolution kernel sizes etc. That doesn't mean that those convolution kernels cannot be applied to images of other sizes. 

For example in your case: 

<pre><code>from keras.models import Model
from keras.layers import Dense,Flatten
from keras.applications import vgg16
from keras import backend as K

model = vgg16.VGG16(weights='imagenet', include_top=False, input_shape=(160,320,3))
model.summary(line_length=150)

flatten = Flatten()
new_layer2 = Dense(10, activation='softmax', name='my_dense_2')

inp2 = model.input
out2 = new_layer2(flatten(model.output))

model2 = Model(inp2, out2)
model2.summary(line_length=150)
</code></pre>

According to <a href="https://github.com/fchollet/keras/blob/master/keras/applications/vgg16.py#L77" rel="noreferrer">here</a> the minimum image size can be <code>48x48x3</code> anything above than that is fine. 

Now its true the original weights were learnt on <code>224,224,3</code> shaped images but the filters weights act as very good starting point for new tasks with new set of images. You do need to re-train the network but the network would converge very quickly. This is the basis of transfer learning.

blocks|key|84142|text|您需要做两件事：|type|unstyled|depth|inlineStyleRanges|entityRanges|data|84143|84144|通过为图像宽度和高度定义None来显式声明输入形状具有可变大小的输入。|ordered-list-item|84145|不使用flatten()，因为它依赖于固定的输入形状。取而代之的是使用GlobalMaxPooling，它不仅可以进行自适应池化，还可以使FC的输入张量扁平化。|84146|84147|我希望这能帮助你实现你想要的东西。|84148|entityMap^0|0|0|0|0|0|0^^$0|@$1|2|3|4|5|6|7|M|8|@]|9|@]|A|$]]|$1|B|3|-4|5|6|7|N|8|@]|9|@]|A|$]]|$1|C|3|D|5|E|7|O|8|@]|9|@]|A|$]]|$1|F|3|G|5|E|7|P|8|@]|9|@]|A|$]]|$1|H|3|-4|5|6|7|Q|8|@]|9|@]|A|$]]|$1|I|3|J|5|6|7|R|8|@]|9|@]|A|$]]|$1|K|3|-4|5|6|7|S|8|@]|9|@]|A|$]]]|L|$]]

There are two things that you need to do:

<ol>
<li>Explicitly declare the input shape to have variable sized inputs by defining None for the image width and height.</li>
<li>Donot use flatten() as it relies on the fixed input shape. Instead use GlobalMaxPooling that will not only do the adaptive pooling, but also flatten the input tensor for the FC to work on.</li>
</ol>

I hope this will help you achieve what you what.

blocks|key|84429|text|可以使用Opencv库的resize()函数。|type|unstyled|depth|inlineStyleRanges|entityRanges|data|84430|+import+cv2
++++width+=+int(224)
++++height+=+int(224)
++++dim+=+(width,+height)
++++'''images+contains+original+dimension+image+array'''
++++resized_images=[]
++++for+i+in+range(0,images.shape[0]):
+++++++++++resized+=+cv2.resize(images[i],+dim,+interpolation+=+cv2.INTER_AREA)
+++++++++++resized_images.append(resized)|code-block|syntax|javascript|84431|entityMap^0|0|0^^$0|@$1|2|3|4|5|6|7|I|8|@]|9|@]|A|$]]|$1|B|3|C|5|D|7|J|8|@]|9|@]|A|$E|F]]|$1|G|3|-4|5|6|7|K|8|@]|9|@]|A|$]]]|H|$]]

You can use resize() function of Opencv library.

<pre><code> import cv2
 width = int(224)
 height = int(224)
 dim = (width, height)
 '''images contains original dimension image array'''
 resized_images=[]
 for i in range(0,images.shape[0]):
 resized = cv2.resize(images[i], dim, interpolation = cv2.INTER_AREA)
 resized_images.append(resized)
</code></pre>

I want to feed images with the shape (160,320,3) to 

<pre><code> VGG16(input_tensor=input_tensor, include_top=False) 
</code></pre>

How can I include a layer that reshapes the images to the shape expected by the VGG16 model, which is (224,224,3) ?

Change input tensor shape for VGG16 application

翻译质量差，导致语言生硬或混乱。

没有提供实际的解决方法或示例。

解答不清晰，无法理解或解决问题。

页面排版不美观，阅读体验差。

文章

问答

视频

学习中心

腾讯云实验室

直播

竞赛

腾讯云代码分析专区

腾讯iOA零信任安全管理系统专区

腾讯云架构师技术同盟交流圈

腾讯云数据库专区

腾讯云顾问专区

腾讯云原生专区

腾讯混元专区

腾讯云TCE专区

腾讯云Lighthouse专区

腾讯云HAI专区

腾讯云Edgeone专区

腾讯云存储专区

腾讯云智能专区

腾讯轻联专区 

腾讯云开发专区

TAPD专区

腾讯轻量云游戏服专区

腾讯云最具价值专家

腾讯云架构师技术同盟

腾讯云创作之星

腾讯云开发者先锋

腾讯云代码助手

云原生构建

TAPD 敏捷项目管理

Cloud Studio

SDK中心

API中心

命令行工具

涵盖代码开发、场景应用、自动测试全流程，助你从零构建专属AI助手

一站式MCP教程库，解锁AI应用新玩法

我想将形状(160,320,3)的图像馈送到 VGG16(input_tensor=input_tensor, include_top=False)   如何包含将图像重塑为VGG16模型所期望的形状的层，即(224,224,3)？

问更改VGG16应用程序的输入张量形状
EN

Stack Overflow用户

社区

活动

圈层

关于

腾讯云开发者

热门产品

热门推荐

更多推荐

问更改VGG16应用程序的输入张量形状EN