Proceedings of the Korean Society of Computer Information Conference (한국컴퓨터정보학회:학술대회논문집)
- 2019.01a
- /
- Pages.123-124
- /
- 2019
An Improved Deep Learning Method for Animal Images
동물 이미지를 위한 향상된 딥러닝 학습
- Wang, Guangxing (School of Com. Inf. & Comm. Eng., Kunsan National University) ;
- Shin, Seong-Yoon (School of Com. Inf. & Comm. Eng., Kunsan National University) ;
- Shin, Kwang-Weong (Dept. of Digital Contents Eng., Wonkwang University) ;
- Lee, Hyun-Chang (Dept. of Digital Contents Eng., Wonkwang University)
- Published : 2019.01.16
Abstract
This paper proposes an improved deep learning method based on small data sets for animal image classification. Firstly, we use a CNN to build a training model for small data sets, and use data augmentation to expand the data samples of the training set. Secondly, using the pre-trained network on large-scale datasets, such as VGG16, the bottleneck features in the small dataset are extracted and to be stored in two NumPy files as new training datasets and test datasets. Finally, training a fully connected network with the new datasets. In this paper, we use Kaggle famous Dogs vs Cats dataset as the experimental dataset, which is a two-category classification dataset.