Conference paper Open Access
Pittaras, Nikiforos; Markatopoulou, Foteini; Mezaris, Vasileios; Patras, Ioannis
In this study we compare three different fine-tuning strategies in order to investigate the best way to transfer the parameters of popular deep convolutional neural networks that were trained for a visual annotation task on one dataset, to a new, considerably different dataset. We focus on the concept-based image/video annotation problem and use ImageNet as the source dataset, while the TRECVID SIN 2013 and PASCAL VOC-2012 classification datasets are used as the target datasets. A large set of experiments examines the effectiveness of three fine-tuning strategies on each of three different pre-trained DCNNs and each target dataset. The reported results give rise to guidelines for effectively fine-tuning a DCNN for concept-based visual annotation.
Name | Size | |
---|---|---|
mmm17_1_preprint.pdf
md5:a0fcfbb9f6eec5eb87e83b1dde238408 |
255.0 kB | Download |
Views | 117 |
Downloads | 84 |
Data volume | 21.4 MB |
Unique views | 117 |
Unique downloads | 81 |