Pluralistic image completion
Most image completion methods produce only one result for each masked input, although there may be many reasonable possibilities. In this paper, we present an approach for pluralistic image completion - the task of generating multiple and diverse plausible solutions for image completion. A major cha...
Saved in:
Main Authors: | Zheng, Chuanxia, Cham, Tat-Jen, Cai, Jianfei |
---|---|
Other Authors: | School of Computer Science and Engineering |
Format: | Conference or Workshop Item |
Language: | English |
Published: |
2020
|
Subjects: | |
Online Access: | https://hdl.handle.net/10356/138481 |
Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
Institution: | Nanyang Technological University |
Language: | English |
Similar Items
-
Pluralistic free-form image completion
by: Zheng, Chuanxia, et al.
Published: (2023) -
Bridging global context interactions for high-fidelity image completion
by: Zheng, Chuanxia, et al.
Published: (2023) -
Recovering facial reflectance and geometry from multi-view images
by: Song, Guoxian, et al.
Published: (2023) -
Visiting the Invisible: layer-by-layer completed scene decomposition
by: Zheng, Chuanxia, et al.
Published: (2023) -
T2Net : synthetic-to-realistic translation for solving single-image depth estimation tasks
by: Zheng, Chuanxia, et al.
Published: (2020)