File Download
There are no files associated with this item.
Supplementary
-
Citations:
- Appears in Collections:
Conference Paper: Pseudo-labeled auto-curriculum learning for semi-supervised keypoint localization
Title | Pseudo-labeled auto-curriculum learning for semi-supervised keypoint localization |
---|---|
Authors | |
Issue Date | 2022 |
Publisher | IEEE. |
Citation | The tenth International Conference on Learning Representation (ICLR) (Virtual), 25-29 April, 2022 How to Cite? |
Abstract | Localizing keypoints of an object is a basic visual problem. However, supervised learning of a keypoint localization network often requires a large amount of data, which is expensive and time-consuming to obtain. To remedy this, there is an ever-growing interest in semi-supervised learning (SSL), which leverages a small set of labeled data along with a large set of unlabeled data. Among these SSL approaches, pseudo-labeling (PL) is one of the most popular. PL approaches apply pseudo-labels to unlabeled data, and then train the model with a combination of the labeled and pseudo-labeled data iteratively. The key to the success of PL is the selection of high-quality pseudo-labeled samples. Previous works mostly select training samples by manually setting a single confidence threshold. We propose to automatically select reliable pseudo-labeled samples with a series of dynamic thresholds, which constitutes a learning curriculum. Extensive experiments on six keypoint localization benchmark datasets demonstrate that the proposed approach significantly outperforms the previous state-of-the-art SSL approaches. |
Persistent Identifier | http://hdl.handle.net/10722/315859 |
DC Field | Value | Language |
---|---|---|
dc.contributor.author | Wang, C | - |
dc.contributor.author | Jin, S | - |
dc.contributor.author | Guan, Y | - |
dc.contributor.author | Liu, W | - |
dc.contributor.author | Qian, C | - |
dc.contributor.author | Luo, P | - |
dc.contributor.author | Ouyang, W | - |
dc.date.accessioned | 2022-08-19T09:05:44Z | - |
dc.date.available | 2022-08-19T09:05:44Z | - |
dc.date.issued | 2022 | - |
dc.identifier.citation | The tenth International Conference on Learning Representation (ICLR) (Virtual), 25-29 April, 2022 | - |
dc.identifier.uri | http://hdl.handle.net/10722/315859 | - |
dc.description.abstract | Localizing keypoints of an object is a basic visual problem. However, supervised learning of a keypoint localization network often requires a large amount of data, which is expensive and time-consuming to obtain. To remedy this, there is an ever-growing interest in semi-supervised learning (SSL), which leverages a small set of labeled data along with a large set of unlabeled data. Among these SSL approaches, pseudo-labeling (PL) is one of the most popular. PL approaches apply pseudo-labels to unlabeled data, and then train the model with a combination of the labeled and pseudo-labeled data iteratively. The key to the success of PL is the selection of high-quality pseudo-labeled samples. Previous works mostly select training samples by manually setting a single confidence threshold. We propose to automatically select reliable pseudo-labeled samples with a series of dynamic thresholds, which constitutes a learning curriculum. Extensive experiments on six keypoint localization benchmark datasets demonstrate that the proposed approach significantly outperforms the previous state-of-the-art SSL approaches. | - |
dc.language | eng | - |
dc.publisher | IEEE. | - |
dc.relation.ispartof | International Conference on Learning Representation (ICLR) | - |
dc.rights | . Copyright © IEEE. | - |
dc.title | Pseudo-labeled auto-curriculum learning for semi-supervised keypoint localization | - |
dc.type | Conference_Paper | - |
dc.identifier.email | Luo, P: pluo@hku.hk | - |
dc.identifier.authority | Luo, P=rp02575 | - |
dc.identifier.doi | 10.48550/arXiv:2201.08613v2 | - |
dc.identifier.hkuros | 335578 | - |
dc.publisher.place | United States | - |