Abstract
Recognizing objects in natural images is an intricate problem involving multiple conflicting objectives. Deep convolutional neural networks, trained on large datasets, achieve convincing results and are currently the state-of-the-art approach for this task. However, the long time needed to train such deep networks is a major drawback. We tackled this problem by reusing a previously trained network. For this purpose, we first trained a deep convolutional network on the ILSVRC-12 dataset. We then maintained the learned convolution kernels and only retrained the classification part on different datasets. Using this approach, we achieved an accuracy of 67.68% on CIFAR-100, compared to the previous state-of-the-art result of 65.43%. Furthermore, our findings indicate that convolutional networks are able to learn generic feature extractors that can be used for different tasks.
| Originalsprache | Englisch |
|---|---|
| Titel | 2015 International Joint Conference on Neural Networks (IJCNN) |
| Seitenumfang | 4 |
| Herausgeber (Verlag) | IEEE |
| Erscheinungsdatum | 28.09.2015 |
| ISBN (Print) | 978-1-4799-1961-1 |
| ISBN (elektronisch) | 978-1-4799-1960-4 |
| DOIs | |
| Publikationsstatus | Veröffentlicht - 28.09.2015 |
| Veranstaltung | International Joint Conference on Neural Networks 2015 - Killarney, Irland Dauer: 12.07.2015 → 17.07.2015 https://ieeexplore.ieee.org/document/7280683/ |
UN SDGs
Dieser Output leistet einen Beitrag zu folgendem(n) Ziel(en) für nachhaltige Entwicklung
-
SDG 3 – Gesundheit und Wohlergehen
-
SDG 9 – Industrie, Innovation und Infrastruktur
Fingerprint
Untersuchen Sie die Forschungsthemen von „Deep Convolutional Neural Networks as Generic Feature Extractors“. Zusammen bilden sie einen einzigartigen Fingerprint.Dieses zitieren
- APA
- Author
- BIBTEX
- Harvard
- Standard
- RIS
- Vancouver