Unsupervised Domain Adaptation for Image Classification and Object Detection Using Guided Transfer Learning Approach and JS Divergence
Unsupervised domain adaptation (UDA) is a transfer learning technique utilized in deep learning. UDA aims to reduce the distribution gap between labeled source and unlabeled target domains by adapting a model through fine-tuning. Typically, UDA approaches assume the same categories in both domains....
| Main Authors: | , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2023-04-01
|
| Series: | Sensors |
| Subjects: | |
| Online Access: | https://www.mdpi.com/1424-8220/23/9/4436 |
| _version_ | 1797601699468148736 |
|---|---|
| author | Parth Goel Amit Ganatra |
| author_facet | Parth Goel Amit Ganatra |
| author_sort | Parth Goel |
| collection | DOAJ |
| description | Unsupervised domain adaptation (UDA) is a transfer learning technique utilized in deep learning. UDA aims to reduce the distribution gap between labeled source and unlabeled target domains by adapting a model through fine-tuning. Typically, UDA approaches assume the same categories in both domains. The effectiveness of transfer learning depends on the degree of similarity between the domains, which determines an efficient fine-tuning strategy. Furthermore, domain-specific tasks generally perform well when the feature distributions of the domains are similar. However, utilizing a trained source model directly in the target domain may not generalize effectively due to domain shift. Domain shift can be caused by intra-class variations, camera sensor variations, background variations, and geographical changes. To address these issues, we design an efficient unsupervised domain adaptation network for image classification and object detection that can learn transferable feature representations and reduce the domain shift problem in a unified network. We propose the guided transfer learning approach to select the layers for fine-tuning the model, which enhances feature transferability and utilizes the JS-Divergence to minimize the domain discrepancy between the domains. We evaluate our proposed approaches using multiple benchmark datasets. Our domain adaptive image classification approach achieves 93.2% accuracy on the Office-31 dataset and 75.3% accuracy on the Office-Home dataset. In addition, our domain adaptive object detection approach achieves 51.1% mAP on the Foggy Cityscapes dataset and 72.7% mAP on the Indian Vehicle dataset. We conduct extensive experiments and ablation studies to demonstrate the effectiveness and efficiency of our work. Experimental results also show that our work significantly outperforms the existing methods. |
| first_indexed | 2024-03-11T04:07:18Z |
| format | Article |
| id | doaj.art-db5d9d79fe7047cf95f002a88871b7f3 |
| institution | Directory Open Access Journal |
| issn | 1424-8220 |
| language | English |
| last_indexed | 2024-03-11T04:07:18Z |
| publishDate | 2023-04-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Sensors |
| spelling | doaj.art-db5d9d79fe7047cf95f002a88871b7f32023-11-17T23:44:22ZengMDPI AGSensors1424-82202023-04-01239443610.3390/s23094436Unsupervised Domain Adaptation for Image Classification and Object Detection Using Guided Transfer Learning Approach and JS DivergenceParth Goel0Amit Ganatra1Computer Science & Engineering Department, Devang Patel Institute of Advance Technology and Research (DEPSTAR), Faculty of Technology & Engineering, Charotar University of Science and Technology (CHARUSAT), Changa 388421, Anand, IndiaComputer Science and Engineering Department, Faculty of Engineering & Technology, Parul University (PU), Waghodia 391760, Vadodara, IndiaUnsupervised domain adaptation (UDA) is a transfer learning technique utilized in deep learning. UDA aims to reduce the distribution gap between labeled source and unlabeled target domains by adapting a model through fine-tuning. Typically, UDA approaches assume the same categories in both domains. The effectiveness of transfer learning depends on the degree of similarity between the domains, which determines an efficient fine-tuning strategy. Furthermore, domain-specific tasks generally perform well when the feature distributions of the domains are similar. However, utilizing a trained source model directly in the target domain may not generalize effectively due to domain shift. Domain shift can be caused by intra-class variations, camera sensor variations, background variations, and geographical changes. To address these issues, we design an efficient unsupervised domain adaptation network for image classification and object detection that can learn transferable feature representations and reduce the domain shift problem in a unified network. We propose the guided transfer learning approach to select the layers for fine-tuning the model, which enhances feature transferability and utilizes the JS-Divergence to minimize the domain discrepancy between the domains. We evaluate our proposed approaches using multiple benchmark datasets. Our domain adaptive image classification approach achieves 93.2% accuracy on the Office-31 dataset and 75.3% accuracy on the Office-Home dataset. In addition, our domain adaptive object detection approach achieves 51.1% mAP on the Foggy Cityscapes dataset and 72.7% mAP on the Indian Vehicle dataset. We conduct extensive experiments and ablation studies to demonstrate the effectiveness and efficiency of our work. Experimental results also show that our work significantly outperforms the existing methods.https://www.mdpi.com/1424-8220/23/9/4436domain adaptationtransfer learningimage classificationobject detectionconvolutional neural networkdeep learning |
| spellingShingle | Parth Goel Amit Ganatra Unsupervised Domain Adaptation for Image Classification and Object Detection Using Guided Transfer Learning Approach and JS Divergence Sensors domain adaptation transfer learning image classification object detection convolutional neural network deep learning |
| title | Unsupervised Domain Adaptation for Image Classification and Object Detection Using Guided Transfer Learning Approach and JS Divergence |
| title_full | Unsupervised Domain Adaptation for Image Classification and Object Detection Using Guided Transfer Learning Approach and JS Divergence |
| title_fullStr | Unsupervised Domain Adaptation for Image Classification and Object Detection Using Guided Transfer Learning Approach and JS Divergence |
| title_full_unstemmed | Unsupervised Domain Adaptation for Image Classification and Object Detection Using Guided Transfer Learning Approach and JS Divergence |
| title_short | Unsupervised Domain Adaptation for Image Classification and Object Detection Using Guided Transfer Learning Approach and JS Divergence |
| title_sort | unsupervised domain adaptation for image classification and object detection using guided transfer learning approach and js divergence |
| topic | domain adaptation transfer learning image classification object detection convolutional neural network deep learning |
| url | https://www.mdpi.com/1424-8220/23/9/4436 |
| work_keys_str_mv | AT parthgoel unsuperviseddomainadaptationforimageclassificationandobjectdetectionusingguidedtransferlearningapproachandjsdivergence AT amitganatra unsuperviseddomainadaptationforimageclassificationandobjectdetectionusingguidedtransferlearningapproachandjsdivergence |