Zobrazeno 1 - 10
of 19
pro vyhledávání: '"Seong Joon Oh"'
Publikováno v:
IEEE Transactions on Pattern Analysis and Machine Intelligence. 45:1732-1748
Weakly-supervised object localization (WSOL) has gained popularity over the last years for its promise to train localization models with only image-level labels. Since the seminal WSOL work of class activation mapping (CAM), the field has focused on
Externí odkaz:
https://explore.openaire.eu/search/publication?articleId=doi_dedup___::0e4f3ca82f649d87429d1f17aa5b2e3d
https://doi.org/10.1109/tpami.2022.3169881
https://doi.org/10.1109/tpami.2022.3169881
Publikováno v:
Lecture Notes in Computer Science ISBN: 9783031200731
Externí odkaz:
https://explore.openaire.eu/search/publication?articleId=doi_________::0af1003822df7ab33d2e96af7e79e2b7
https://doi.org/10.1007/978-3-031-20074-8_1
https://doi.org/10.1007/978-3-031-20074-8_1
Publikováno v:
CVPR
Cross-modal retrieval methods build a common representation space for samples from multiple modalities, typically from the vision and the language domains. For images and their captions, the multiplicity of the correspondences makes the task particul
Externí odkaz:
https://explore.openaire.eu/search/publication?articleId=doi_dedup___::f353bcc9710fd227b28caa558898be34
http://arxiv.org/abs/2101.05068
http://arxiv.org/abs/2101.05068
Publikováno v:
CVPR
ImageNet has been arguably the most popular image classification benchmark, but it is also the one with a significant level of label noise. Recent studies have shown that many samples contain multiple classes, despite being assumed to be a single-lab
Externí odkaz:
https://explore.openaire.eu/search/publication?articleId=doi_dedup___::39f311b10f5c0b9873e32cb3b530c8ea
http://arxiv.org/abs/2101.05022
http://arxiv.org/abs/2101.05022
Publikováno v:
CVPR
Weakly-supervised object localization (WSOL) has gained popularity over the last years for its promise to train localization models with only image-level labels. Since the seminal WSOL work of class activation mapping (CAM), the field has focused on
Externí odkaz:
https://explore.openaire.eu/search/publication?articleId=doi_dedup___::b4e293c3bd58aa21b62df0b4a7fb61bd
Publikováno v:
Pattern Recognition. 116:107949
Weakly supervised object localization (WSOL) methods utilize the internal feature responses of a classifier trained only on image-level labels. Classifiers tend to focus on the most discriminative part of the target object, instead of considering its
Externí odkaz:
https://explore.openaire.eu/search/publication?articleId=doi_________::74ed5a70823467ff75d319ec1edfe98d
https://doi.org/10.1016/j.patcog.2021.107949
https://doi.org/10.1016/j.patcog.2021.107949
Publikováno v:
ICCV
Regional dropout strategies have been proposed to enhance the performance of convolutional neural network classifiers. They have proved to be effective for guiding the model to attend on less discriminative parts of objects (e.g. leg as opposed to he
Externí odkaz:
https://explore.openaire.eu/search/publication?articleId=doi_dedup___::7d053dbf96d7e5775854393e7334a134
http://arxiv.org/abs/1905.04899
http://arxiv.org/abs/1905.04899
Autor:
Sungrae Park, Geewook Kim, Junyeop Lee, Sangdoo Yun, Jeonghun Baek, Hwalsuk Lee, Dongyoon Han, Seong Joon Oh
Publikováno v:
ICCV
Many new proposals for scene text recognition (STR) models have been introduced in recent years. While each claim to have pushed the boundary of the technology, a holistic and fair comparison has been largely missing in the field due to the inconsist
Externí odkaz:
https://explore.openaire.eu/search/publication?articleId=doi_dedup___::a937edce67174924ecb30a073b43f0b8
http://arxiv.org/abs/1904.01906
http://arxiv.org/abs/1904.01906
Publikováno v:
CVPR Workshops
Scene text recognition (STR) is the task of recognizing character sequences in natural scenes. While there have been great advances in STR methods, current methods which convert two-dimensional (2D) image to one-dimensional (1D) feature map still fai
Externí odkaz:
https://explore.openaire.eu/search/publication?articleId=doi_dedup___::85a7acb83b0f22e3d01d26a4ff4cd814
Publikováno v:
Explainable AI: Interpreting, Explaining and Visualizing Deep Learning ISBN: 9783030289539
ICLR (Poster)
ICLR (Poster)
Much progress in interpretable AI is built around scenarios where the user, one who interprets the model, has a full ownership of the model to be diagnosed. The user either owns the training data and computing resources to train an interpretable mode
Externí odkaz:
https://explore.openaire.eu/search/publication?articleId=doi_________::90a0015bfe7f0f734ddd3eaa50bd7c5d
https://doi.org/10.1007/978-3-030-28954-6_7
https://doi.org/10.1007/978-3-030-28954-6_7