Deep Convolutional Neural Networks as Generic Feature Extractors
This work addresses training efficiency for computer vision researchers, but it is incremental as it builds on existing transfer learning concepts.
The paper tackled the problem of long training times for deep convolutional neural networks by reusing a previously trained network's convolution kernels and retraining only the classification part on different datasets, achieving an accuracy of 67.68% on CIFAR-100 compared to the previous state-of-the-art of 65.43%.
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 ILSVRC2012 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.