How can I add additional features to a pretrained AlexNet?
4 views (last 30 days)
Show older comments
MathWorks Support Team
on 25 Jan 2018
Edited: MathWorks Support Team
on 26 Aug 2021
I am classifying images using transfer learning and the pretrained AlexNet convolutional neural network.
I would like to add additional features in the fully connected layers at the end of the network.
Is it possible to add features that bypass the convolutional layers and are incorporated only at the fully connected layers?
Accepted Answer
MathWorks Support Team
on 26 Aug 2021
Edited: MathWorks Support Team
on 26 Aug 2021
You can do this by following the workflow:
1. Extract the features that is outputted by the pretrained (& transfer learning) AlexNet network right before the fully connected layer:
2. Concatenate the features from step 1 with the additional features from your image metadata.
3. Create and train a fully-connected neural network that will take in the concatenated features from step 2.
0 Comments
More Answers (0)
See Also
Categories
Find more on Image Data Workflows in Help Center and File Exchange
Community Treasure Hunt
Find the treasures in MATLAB Central and discover how the community can help you!
Start Hunting!