A study of CNN transfer learning for image processing
Transfer learning, a domain of machine learning, seeks to be an efficient solution over traditional machine learning techniques by adapting existing convolutional neural networks (CNN) to suit a new problem. Adapting a CNN for transfer learning can be done through the changing of hyperparameters and...
| Main Author: | |
|---|---|
| Other Authors: | |
| Format: | Final Year Project (FYP) |
| Language: | English |
| Published: |
Nanyang Technological University
2020
|
| Subjects: | |
| Online Access: | https://hdl.handle.net/10356/145039 |
| _version_ | 1826128665752109056 |
|---|---|
| author | Koh, Yee Zuo |
| author2 | Kai-Kuang Ma |
| author_facet | Kai-Kuang Ma Koh, Yee Zuo |
| author_sort | Koh, Yee Zuo |
| collection | NTU |
| description | Transfer learning, a domain of machine learning, seeks to be an efficient solution over traditional machine learning techniques by adapting existing convolutional neural networks (CNN) to suit a new problem. Adapting a CNN for transfer learning can be done through the changing of hyperparameters and the freezing of CNN’s layers.
In this paper, transfer learning was implemented to VGG-Face, a state-of-the-art facial recognition CNN, where it was adapted to understand and classify images from the JAFFE dataset consisting of four different human facial emotions: (1) Angry, (2) Happy, (3) Sad, (4) Surprised.
A cascade transfer learning was performed using the FER2013 dataset for the first fine- tune and a portion of the Japanese Female Facial Expression (JAFFE) dataset for the second fine-tune. The test accuracy was then taken using a portion of the JAFFE dataset. The changing of hyperparameters and the freezing of the CNN’s layers within the VGG- Face CNN were also discussed in this paper. The experiments were ran using a NVIDIA RTX 2060 GPU on MATLAB R2020a using its various toolboxes.
The final architecture proposed a validation accuracy of 62.41% on the FER2013 dataset, and a test accuracy 86.11% on the JAFFE test dataset, which was an increase compared to the baseline of 20.63% and 27.78% respectively. |
| first_indexed | 2024-10-01T07:28:40Z |
| format | Final Year Project (FYP) |
| id | ntu-10356/145039 |
| institution | Nanyang Technological University |
| language | English |
| last_indexed | 2024-10-01T07:28:40Z |
| publishDate | 2020 |
| publisher | Nanyang Technological University |
| record_format | dspace |
| spelling | ntu-10356/1450392023-07-07T18:13:11Z A study of CNN transfer learning for image processing Koh, Yee Zuo Kai-Kuang Ma School of Electrical and Electronic Engineering EKKMA@ntu.edu.sg Engineering::Computer science and engineering::Computing methodologies::Image processing and computer vision Transfer learning, a domain of machine learning, seeks to be an efficient solution over traditional machine learning techniques by adapting existing convolutional neural networks (CNN) to suit a new problem. Adapting a CNN for transfer learning can be done through the changing of hyperparameters and the freezing of CNN’s layers. In this paper, transfer learning was implemented to VGG-Face, a state-of-the-art facial recognition CNN, where it was adapted to understand and classify images from the JAFFE dataset consisting of four different human facial emotions: (1) Angry, (2) Happy, (3) Sad, (4) Surprised. A cascade transfer learning was performed using the FER2013 dataset for the first fine- tune and a portion of the Japanese Female Facial Expression (JAFFE) dataset for the second fine-tune. The test accuracy was then taken using a portion of the JAFFE dataset. The changing of hyperparameters and the freezing of the CNN’s layers within the VGG- Face CNN were also discussed in this paper. The experiments were ran using a NVIDIA RTX 2060 GPU on MATLAB R2020a using its various toolboxes. The final architecture proposed a validation accuracy of 62.41% on the FER2013 dataset, and a test accuracy 86.11% on the JAFFE test dataset, which was an increase compared to the baseline of 20.63% and 27.78% respectively. Bachelor of Engineering (Information Engineering and Media) 2020-12-09T05:32:05Z 2020-12-09T05:32:05Z 2020 Final Year Project (FYP) https://hdl.handle.net/10356/145039 en A3331-192 application/pdf Nanyang Technological University |
| spellingShingle | Engineering::Computer science and engineering::Computing methodologies::Image processing and computer vision Koh, Yee Zuo A study of CNN transfer learning for image processing |
| title | A study of CNN transfer learning for image processing |
| title_full | A study of CNN transfer learning for image processing |
| title_fullStr | A study of CNN transfer learning for image processing |
| title_full_unstemmed | A study of CNN transfer learning for image processing |
| title_short | A study of CNN transfer learning for image processing |
| title_sort | study of cnn transfer learning for image processing |
| topic | Engineering::Computer science and engineering::Computing methodologies::Image processing and computer vision |
| url | https://hdl.handle.net/10356/145039 |
| work_keys_str_mv | AT kohyeezuo astudyofcnntransferlearningforimageprocessing AT kohyeezuo studyofcnntransferlearningforimageprocessing |