Julianne Tan
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Julianne Tan has a total of 8 co-authors. Here are the top co-authors by number of shared publications:
- Kai-Lung Hua
National Taiwan University of Science and Technology
2 shared publications - Tai-Ming Huang
National Taiwan University, Taipei, Taiwan and Academia Sinica, Taipei, Taiwan
2 shared publications - Jilyan Dy
Department of Computer Science and Information Engineering, National Taiwan University of Science and Technology, Taipei, Taiwan
1 shared publication Shang-Che Hsu
National Taiwan University of Science and Technology, Taipei, Taiwan
1 shared publication- Shang-Fu Chen
Scripps Research Translational Institute, La Jolla, USA
1 shared publication Ting-Kai Yang
Grad. Inst. Automation Technology, National Taipei University of Technology,Taipei,Taiwan,106
1 shared publication
Latest Publications
- 6 July 2022
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 407-408
A Data Hiding Scheme Based On Absolute Moment Block Truncation Coding and Lookup Table
This research proposes a data hiding scheme which uses absolute moment block truncation coding (AMBTC) for image compression, and generates a reference lookup table which has a one-to-many relationship between the hidden data and the candidate bitmaps of the compressed image. When embedding data, the candidate blocks that match the hidden data are filtered through mean square error (MSE) function to select the optimal block. From the preliminary experiments, we prove that the proposed method is superior to other data hiding method in terms of data hiding ability and image visual quality.
- 6 July 2022
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 407-408
A Data Hiding Scheme Based On Absolute Moment Block Truncation Coding and Lookup Table
This research proposes a data hiding scheme which uses absolute moment block truncation coding (AMBTC) for image compression, and generates a reference lookup table which has a one-to-many relationship between the hidden data and the candidate bitmaps of the compressed image. When embedding data, the candidate blocks that match the hidden data are filtered through mean square error (MSE) function to select the optimal block. From the preliminary experiments, we prove that the proposed method is superior to other data hiding method in terms of data hiding ability and image visual quality.
- 24 May 2022
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE MultiMedia
- Vol. 29 (4) , 59-68
Unpaired Image-to-Image Translation Using Negative Learning for Noisy Patches
Unpaired image-to-image translation finds a mapping between two domains that do not have paired data. One approach is patchwise contrastive learning, a one-sided translation that maximizes mutual information between corresponding input and output patches. Non-corresponding patches are treated as negatives. Previous approaches randomly select non-corresponding patches, resulting in semantically similar patches incorrectly labeled as negatives. Inspired by negative learning, we propose the novel patch-wise negative learning loss to address this issue. We do not naively minimize mutual information between all non-corresponding ones, unlike prior methods. Instead, we choose one non-corresponding patch and maximize dissimilarity with the query patch. The selected non-corresponding patch reduces the chance of choosing false negatives that contain high mutual information. By further maximizing dissimilarity with that single negative, we discourage our model from fitting on noisy negative patches. We demonstrate the capabilities of our model against other prominent image translation methods on the selfie2anime, horse2zebra, and cat2dog datasets.
- 24 May 2022
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE MultiMedia
- Vol. 29 (4) , 59-68
Unpaired Image-to-Image Translation Using Negative Learning for Noisy Patches
Unpaired image-to-image translation finds a mapping between two domains that do not have paired data. One approach is patchwise contrastive learning, a one-sided translation that maximizes mutual information between corresponding input and output patches. Non-corresponding patches are treated as negatives. Previous approaches randomly select non-corresponding patches, resulting in semantically similar patches incorrectly labeled as negatives. Inspired by negative learning, we propose the novel patch-wise negative learning loss to address this issue. We do not naively minimize mutual information between all non-corresponding ones, unlike prior methods. Instead, we choose one non-corresponding patch and maximize dissimilarity with the query patch. The selected non-corresponding patch reduces the chance of choosing false negatives that contain high mutual information. By further maximizing dissimilarity with that single negative, we discourage our model from fitting on noisy negative patches. We demonstrate the capabilities of our model against other prominent image translation methods on the selfie2anime, horse2zebra, and cat2dog datasets.