Merge pull request #2 from ImIntheMiddle/hugoblox-import-publications

Hugo Blox Builder - Import latest publications

content/publication/taketsugu-active-2023/cite.bib

@@ -0,0 +1,14 @@
+@misc{taketsugu_active_2023,
+ abstract = {Human Pose (HP) estimation is actively researched because of its wide range of applications. However, even estimators pre-trained on large datasets may not perform satisfactorily due to a domain gap between the training and test data. To address this issue, we present our approach combining Active Learning (AL) and Transfer Learning (TL) to adapt HP estimators to individual video domains efficiently. For efficient learning, our approach quantifies (i) the estimation uncertainty based on the temporal changes in the estimated heatmaps and (ii) the unnaturalness in the estimated full-body HPs. These quantified criteria are then effectively combined with the state-of-the-art representativeness criterion to select uncertain and diverse samples for efficient HP estimator learning. Furthermore, we reconsider the existing Active Transfer Learning (ATL) method to introduce novel ideas related to the retraining methods and Stopping Criteria (SC). Experimental results demonstrate that our method enhances learning efficiency and outperforms comparative methods. Our code is publicly available at: https://github.com/ImIntheMiddle/VATL4Pose-WACV2024},
+ annote = {Comment: 17 pages, 12 figures, Accepted by WACV 2024},
+ author = {Taketsugu, Hiromu and Ukita, Norimichi},
+ copyright = {All rights reserved},
+ doi = {10.48550/arXiv.2311.05041},
+ month = {November},
+ note = {arXiv:2311.05041 [cs]},
+ publisher = {arXiv},
+ title = {Active Transfer Learning for Efficient Video-Specific Human Pose Estimation},
+ url = {http://arxiv.org/abs/2311.05041},
+ urldate = {2023-11-24},
+ year = {2023}
+}

content/publication/taketsugu-active-2023/index.md

@@ -0,0 +1,29 @@
+---
+title: Active Transfer Learning for Efficient Video-Specific Human Pose Estimation
+authors:
+- Hiromu Taketsugu
+- Norimichi Ukita
+date: '2023-11-01'
+publishDate: '2023-11-24T21:29:26.969043Z'
+publication_types:
+- manuscript
+publication: '*arXiv*'
+doi: 10.48550/arXiv.2311.05041
+abstract: 'Human Pose (HP) estimation is actively researched because of its wide range
+  of applications. However, even estimators pre-trained on large datasets may not
+  perform satisfactorily due to a domain gap between the training and test data. To
+  address this issue, we present our approach combining Active Learning (AL) and Transfer
+  Learning (TL) to adapt HP estimators to individual video domains efficiently. For
+  efficient learning, our approach quantifies (i) the estimation uncertainty based
+  on the temporal changes in the estimated heatmaps and (ii) the unnaturalness in
+  the estimated full-body HPs. These quantified criteria are then effectively combined
+  with the state-of-the-art representativeness criterion to select uncertain and diverse
+  samples for efficient HP estimator learning. Furthermore, we reconsider the existing
+  Active Transfer Learning (ATL) method to introduce novel ideas related to the retraining
+  methods and Stopping Criteria (SC). Experimental results demonstrate that our method
+  enhances learning efficiency and outperforms comparative methods. Our code is publicly
+  available at: https://github.com/ImIntheMiddle/VATL4Pose-WACV2024'
+links:
+- name: URL
+  url: http://arxiv.org/abs/2311.05041
+---

content/publication/taketsugu-uncertainty-2023/cite.bib

@@ -0,0 +1,13 @@
+@inproceedings{taketsugu_uncertainty_2023,
+ abstract = {This paper presents a combination of Active Learning (AL) and Transfer Learning (TL) for efficiently adapting Human Pose (HP) estimators to individual videos. The proposed approach quantifies estimation uncertainty through the temporal changes and unnaturalness of estimated HPs. These uncertainty criteria are combined with clustering-based representativeness criterion to avoid the useless selection of similar samples. Experiments demonstrated that the proposed method achieves high learning efficiency and outperforms comparative methods.},
+ author = {Taketsugu, Hiromu and Ukita, Norimichi},
+ booktitle = {2023 18th International Conference on Machine Vision and Applications (MVA)},
+ copyright = {All rights reserved},
+ doi = {10.23919/MVA57639.2023.10215565},
+ month = {July},
+ pages = {1--5},
+ title = {Uncertainty Criteria in Active Transfer Learning for Efficient Video-Specific Human Pose Estimation},
+ url = {https://ieeexplore.ieee.org/abstract/document/10215565},
+ urldate = {2023-11-24},
+ year = {2023}
+}

content/publication/taketsugu-uncertainty-2023/index.md

@@ -0,0 +1,24 @@
+---
+title: Uncertainty Criteria in Active Transfer Learning for Efficient Video-Specific
+  Human Pose Estimation
+authors:
+- Hiromu Taketsugu
+- Norimichi Ukita
+date: '2023-07-01'
+publishDate: '2023-11-24T21:29:26.986148Z'
+publication_types:
+- paper-conference
+publication: '*2023 18th International Conference on Machine Vision and Applications
+  (MVA)*'
+doi: 10.23919/MVA57639.2023.10215565
+abstract: This paper presents a combination of Active Learning (AL) and Transfer Learning
+  (TL) for efficiently adapting Human Pose (HP) estimators to individual videos. The
+  proposed approach quantifies estimation uncertainty through the temporal changes
+  and unnaturalness of estimated HPs. These uncertainty criteria are combined with
+  clustering-based representativeness criterion to avoid the useless selection of
+  similar samples. Experiments demonstrated that the proposed method achieves high
+  learning efficiency and outperforms comparative methods.
+links:
+- name: URL
+  url: https://ieeexplore.ieee.org/abstract/document/10215565
+---