清空筛选条件
清空筛选条件

Augmented image data store wont work unless i write Augmented image Source, is this stopping me from making a data store?

1 次查看（过去 30 天）
Nathan Reilly
Nathan Reilly 2018-4-9
回答： Helen Victoria 2018-12-14
Compile errors says:
Error using trainingOptions (line 187)
The value of 'ValidationData' is invalid. Validation data must be a table, an ImageDatastore, or a cell
array with input data and responses.

请先登录，再回答此问题。

回答（2 个）

Walter Roberson
Walter Roberson 2018-4-11
augmented image data store was not introduced until R2018a.
  2 个评论
Nathan Reilly
Nathan Reilly 2018-4-11
I know so everyone says use AugmentedImageSource but when i try to train with my data it says it must be from a data store which my matlab doesnt have. So i have tried using imresize for example: inputSize = [227 227 3]; imds.ReadFcn = @(loc)imresize(imread(loc),[inputSize]); Plus many combinations of this code and more. Still wont work compile error says all images must be the same size. Or it says target images are 227 227 1 but need to be 227 227 3 for alexnet, so i need to use 4Darray but cant get that to work either...
Walter Roberson
Walter Roberson 2018-4-11
The input images must be RGB for alexnet. You can construct RGB images from grayscale images as
imds.ReadFcn = @(loc) repmat(imresize(imread(loc),[inputSize]),[1 1 3]);

请先登录，再进行评论。

Helen Victoria
Helen Victoria 2018-12-14
Error using trainingOptions (line 262)
The value of 'ValidationData' is invalid. MiniBatchable Datastore has no responses.
Error in cifarpretrained (line 24)
options = trainingOptions('sgdm', . Am getting this error while training on alexnet using transfer.I am using 2018b version

类别

在 Help Center 和 File Exchange 中查找有关 Deep Learning for Image Processing 的更多信息

标签

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!

Translated by