Loading...
Please wait, while we are loading the content...
Similar Documents
Fully Convolutional Neural Network
| Content Provider | Scilit |
|---|---|
| Author | Cresson, Rémi |
| Copyright Year | 2020 |
| Description | This chapter introduces the Fully Convolutional Neural Networks (FC) that are just Convolutional Neural Networks (CNNs) that can process entire image regions instead of being limited to small patches. The costliest operations implemented in the deep net are the convolutions, which are massively parallel and can be implemented so that the intermediary results are reused for other patch processing. However, the patch-based mode disables the possible reuse of the intermediate results since they are re-computed at each new patch, which is not efficient. A fully convolutional network runs faster than patch based, enough to process the full image in reasonable time to test it, just remove the extended filename in the previous command. The proposed FCN has a lot more weight than the previously introduced CNN one. One can then expect that the gradient descent will take more time. Book Name: Deep Learning for Remote Sensing Images with Open Source Software |
| Related Links | https://content.taylorfrancis.com/books/download?dac=C2019-0-09527-7&isbn=9781003020851&doi=10.1201/9781003020851-6&format=pdf |
| Ending Page | 50 |
| Page Count | 6 |
| Starting Page | 45 |
| DOI | 10.1201/9781003020851-6 |
| Language | English |
| Publisher | Informa UK Limited |
| Publisher Date | 2020-07-15 |
| Access Restriction | Open |
| Subject Keyword | Book Name: Deep Learning for Remote Sensing Images with Open Source Software Artificial Intelligence Convolutional Neural Networks Cnn Reuse Patch Introduced Fully Convolutional |
| Content Type | Text |
| Resource Type | Chapter |
