Modify a Convolutional Neural Network

Hello,
I am working on a task related to image classification. I want to modify my CNN and add a HOG feature extractor in parallel with the convolution layers extraction process and at the end concatenate all the feature and apply the training of the network:
trainNetwork(augmentedTrainingSet, network, opts);
To apply this modification and get a new CNN model that could be trained end-to-end, I need a new layer that will apply the HOG extraction.
How can I create a layer that will extract, which i can add to the CNN model ? :
features = extractHOGFeatures(I);
I don't know if my idea is clear but I attached a picture explaining the concept. I will appreciate any kind of help!
Thank you in advance,

回答 (0 件)

カテゴリ

ヘルプ センター および File ExchangeDeep Learning Toolbox についてさらに検索

質問済み:

2021 年 5 月 21 日

編集済み:

2021 年 5 月 23 日

Community Treasure Hunt

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

Start Hunting!

Translated by