From b4f50a7ac5366ea63614fc42db7aedb9a2060903 Mon Sep 17 00:00:00 2001 From: Qingyun <79644233+Li-Qingyun@users.noreply.github.com> Date: Mon, 16 Oct 2023 21:58:36 +0800 Subject: [PATCH 1/5] Update README.md of h2rboxv2 --- configs/h2rbox_v2/README.md | 16 +++++++--------- 1 file changed, 7 insertions(+), 9 deletions(-) diff --git a/configs/h2rbox_v2/README.md b/configs/h2rbox_v2/README.md index b2b1e3730..4672ff17a 100644 --- a/configs/h2rbox_v2/README.md +++ b/configs/h2rbox_v2/README.md @@ -1,6 +1,6 @@ # H2RBox-v2 -> [H2RBox-v2: Boosting HBox-supervised Oriented Object Detection via Symmetric Learning](https://arxiv.org/pdf/2304.04403) +> [H2RBox-v2: Incorporating Symmetry for Boosting Horizontal Box Supervised Oriented Object Detection](https://arxiv.org/pdf/2304.04403) @@ -10,7 +10,7 @@ -With the increasing demand for oriented object detection e.g. in autonomous driving and remote sensing, the oriented annotation has become a labor-intensive work. To make full use of existing horizontally annotated datasets and reduce the annotation cost, a weakly-supervised detector H2RBox for learning the rotated box (RBox) from the horizontal box (HBox) has been proposed and received great attention. This paper presents a new version, H2RBox-v2, to further bridge the gap between HBox-supervised and RBox-supervised oriented object detection. While exploiting axisymmetry via flipping and rotating consistencies is available through our theoretical analysis, H2RBox-v2, using a weakly-supervised branch similar to H2RBox, is embedded with a novel self-supervised branch that learns orientations from the symmetry inherent in the image of objects. Complemented by modules to cope with peripheral issues, e.g. angular periodicity, a stable and effective solution is achieved. To our knowledge, H2RBox-v2 is the first symmetry-supervised paradigm for oriented object detection. Compared to H2RBox, our method is less susceptible to low annotation quality and insufficient training data, which in such cases is expected to give a competitive performance much closer to fully-supervised oriented object detectors. Specifically, the performance comparison between H2RBox-v2 and Rotated FCOS on DOTA-v1.0/1.5/2.0 is 72.31%/64.76%/50.33% vs. 72.44%/64.53%/51.77%, 89.66% vs. 88.99% on HRSC, and 42.27% vs. 41.25% on FAIR1M. +With the rapidly increasing demand for oriented object detection e.g. in autonomous driving and remote sensing, the recently proposed paradigm involving weakly-supervised detector H2RBox for learning rotated box (RBox) from the (currently) more readily-available horizontal box (HBox) has shown promise. This paper presents H2RBox-v2, to further bridge the gap between HBox-supervised and RBox-supervised oriented object detection. Specifically, we propose to leverage the reflection symmetry via flip and rotate consistencies, using a weakly-supervised network branch similar to H2RBox, together with a novel self-supervised branch that learns orientations from the symmetry inherent in visual objects. The detector is further stabilized and enhanced by practical techniques to cope with peripheral issues e.g. angular periodicity. To our best knowledge, H2RBox-v2 is the first symmetry-aware self-supervised paradigm for oriented object detection. In particular, our method shows less susceptibility to low-quality annotation and insufficient training data compared to H2RBox. Specifically, H2RBox-v2 achieves very close performance to a rotation annotation trained counterpart -- Rotated FCOS: 1) DOTA-v1.0/1.5/2.0: 72.31%/64.76%/50.33% vs. 72.44%/64.53%/51.77%; 2) HRSC: 89.66% vs. 88.99%; 3) FAIR1M: 42.27% vs. 41.25%. The source code will be made publicly available. ## Results and models @@ -43,12 +43,10 @@ HRSC ## Citation ``` -@misc{yu2023h2rboxv2, -title={H2RBox-v2: Boosting HBox-supervised Oriented Object Detection via Symmetric Learning}, -author={Yi Yu and Xue Yang and Qingyun Li and Yue Zhou and Gefan Zhang and Feipeng Da and Junchi Yan}, -year={2023}, -eprint={2304.04403}, -archivePrefix={arXiv}, -primaryClass={cs.CV} +@inproceedings{yu2023h2rboxv2, + title={H2RBox-v2: Incorporating Symmetry for Boosting Horizontal Box Supervised Oriented Object Detection}, + author={Yi Yu and Xue Yang and Qingyun Li and Yue Zhou and Gefan Zhang and Feipeng Da and Junchi Yan}, + year={2023}, + booktitle={Advances in Neural Information Processing Systems} } ``` From fabef42b4c0875ff1a10bef073479f4705cfa072 Mon Sep 17 00:00:00 2001 From: Qingyun <79644233+Li-Qingyun@users.noreply.github.com> Date: Mon, 16 Oct 2023 22:01:09 +0800 Subject: [PATCH 2/5] Update homepage --- README.md | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/README.md b/README.md index 2ce0dc6ad..dde36eeae 100644 --- a/README.md +++ b/README.md @@ -172,7 +172,7 @@ A summary can be found in the [Model Zoo](docs/en/model_zoo.md) page. - [x] [H2RBox](configs/h2rbox/README.md) (ICLR'2023) - [x] [PSC](configs/psc/README.md) (CVPR'2023) - [x] [RTMDet](configs/rotated_rtmdet/README.md) (arXiv) -- [x] [H2RBox-v2](configs/h2rbox_v2/README.md) (arXiv) +- [x] [H2RBox-v2](configs/h2rbox_v2/README.md) (NeurIPS'2023) From 58c069b126ec6c15e5eb29e4a826a22e3d51eeb3 Mon Sep 17 00:00:00 2001 From: Qingyun <79644233+Li-Qingyun@users.noreply.github.com> Date: Mon, 16 Oct 2023 22:04:42 +0800 Subject: [PATCH 3/5] Update abs --- configs/h2rbox_v2/README.md | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/configs/h2rbox_v2/README.md b/configs/h2rbox_v2/README.md index 4672ff17a..25fc61002 100644 --- a/configs/h2rbox_v2/README.md +++ b/configs/h2rbox_v2/README.md @@ -10,7 +10,7 @@ -With the rapidly increasing demand for oriented object detection e.g. in autonomous driving and remote sensing, the recently proposed paradigm involving weakly-supervised detector H2RBox for learning rotated box (RBox) from the (currently) more readily-available horizontal box (HBox) has shown promise. This paper presents H2RBox-v2, to further bridge the gap between HBox-supervised and RBox-supervised oriented object detection. Specifically, we propose to leverage the reflection symmetry via flip and rotate consistencies, using a weakly-supervised network branch similar to H2RBox, together with a novel self-supervised branch that learns orientations from the symmetry inherent in visual objects. The detector is further stabilized and enhanced by practical techniques to cope with peripheral issues e.g. angular periodicity. To our best knowledge, H2RBox-v2 is the first symmetry-aware self-supervised paradigm for oriented object detection. In particular, our method shows less susceptibility to low-quality annotation and insufficient training data compared to H2RBox. Specifically, H2RBox-v2 achieves very close performance to a rotation annotation trained counterpart -- Rotated FCOS: 1) DOTA-v1.0/1.5/2.0: 72.31%/64.76%/50.33% vs. 72.44%/64.53%/51.77%; 2) HRSC: 89.66% vs. 88.99%; 3) FAIR1M: 42.27% vs. 41.25%. The source code will be made publicly available. +With the rapidly increasing demand for oriented object detection e.g. in autonomous driving and remote sensing, the recently proposed paradigm involving weaklysupervised detector H2RBox for learning rotated box (RBox) from the (currently) more readily-available horizontal box (HBox) has shown promise. This paper presents H2RBox-v2, to further bridge the gap between HBox-supervised and RBox-supervised oriented object detection. Specifically, we propose to leverage the reflection symmetry via flip and rotate consistencies, using a weakly-supervised network branch similar to H2RBox, together with a novel self-supervised branch that learns orientations from the symmetry inherent in visual objects. The detector is further stabilized and enhanced by practical techniques to cope with peripheral issues e.g. angular periodicity. To our best knowledge, H2RBox-v2 is the first symmetry-aware self-supervised paradigm for oriented object detection. In particular, our method shows less susceptibility to low-quality annotation and insufficient training data compared to H2RBox. Specifically, H2RBox-v2 achieves very close performance to a rotation annotation trained counterpart – Rotated FCOS: 1) DOTA-v1.0/1.5/2.0: 72.31%/64.76%/50.33% vs. 72.44%/64.53%/51.77%; 2) HRSC: 89.66% vs. 88.99%; 3) FAIR1M: 42.27% vs. 41.25%. ## Results and models From 9f91d91d741329db9c18804434b2907f07287651 Mon Sep 17 00:00:00 2001 From: Qingyun <79644233+Li-Qingyun@users.noreply.github.com> Date: Mon, 16 Oct 2023 22:06:30 +0800 Subject: [PATCH 4/5] Update abs --- configs/h2rbox_v2/README.md | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/configs/h2rbox_v2/README.md b/configs/h2rbox_v2/README.md index 25fc61002..e21297767 100644 --- a/configs/h2rbox_v2/README.md +++ b/configs/h2rbox_v2/README.md @@ -10,7 +10,7 @@ -With the rapidly increasing demand for oriented object detection e.g. in autonomous driving and remote sensing, the recently proposed paradigm involving weaklysupervised detector H2RBox for learning rotated box (RBox) from the (currently) more readily-available horizontal box (HBox) has shown promise. This paper presents H2RBox-v2, to further bridge the gap between HBox-supervised and RBox-supervised oriented object detection. Specifically, we propose to leverage the reflection symmetry via flip and rotate consistencies, using a weakly-supervised network branch similar to H2RBox, together with a novel self-supervised branch that learns orientations from the symmetry inherent in visual objects. The detector is further stabilized and enhanced by practical techniques to cope with peripheral issues e.g. angular periodicity. To our best knowledge, H2RBox-v2 is the first symmetry-aware self-supervised paradigm for oriented object detection. In particular, our method shows less susceptibility to low-quality annotation and insufficient training data compared to H2RBox. Specifically, H2RBox-v2 achieves very close performance to a rotation annotation trained counterpart – Rotated FCOS: 1) DOTA-v1.0/1.5/2.0: 72.31%/64.76%/50.33% vs. 72.44%/64.53%/51.77%; 2) HRSC: 89.66% vs. 88.99%; 3) FAIR1M: 42.27% vs. 41.25%. +With the rapidly increasing demand for oriented object detection, e.g. in autonomous driving and remote sensing, the recently proposed paradigm involving weakly-supervised detector H2RBox for learning rotated box (RBox) from the more readily-available horizontal box (HBox) has shown promise. This paper presents H2RBox-v2, to further bridge the gap between HBox-supervised and RBox-supervised oriented object detection. Specifically, we propose to leverage the reflection symmetry via flip and rotate consistencies, using a weakly-supervised network branch similar to H2RBox, together with a novel self-supervised branch that learns orientations from the symmetry inherent in visual objects. The detector is further stabilized and enhanced by practical techniques to cope with peripheral issues e.g. angular periodicity. To our best knowledge, H2RBox-v2 is the first symmetry-aware self-supervised paradigm for oriented object detection. In particular, our method shows less susceptibility to low-quality annotation and insufficient training data compared to H2RBox. Specifically, H2RBox-v2 achieves very close performance to a rotation annotation trained counterpart -- Rotated FCOS: 1) DOTA-v1.0/1.5/2.0: 72.31%/64.76%/50.33% vs. 72.44%/64.53%/51.77%; 2) HRSC: 89.66% vs. 88.99%; 3) FAIR1M: 42.27% vs. 41.25%. ## Results and models From ec80524f031b7c3777fe11e4050205ec8a429924 Mon Sep 17 00:00:00 2001 From: yuyi1005 Date: Fri, 13 Sep 2024 13:33:49 +0800 Subject: [PATCH 5/5] Update README.md --- configs/h2rbox_v2/README.md | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/configs/h2rbox_v2/README.md b/configs/h2rbox_v2/README.md index e21297767..dfa93d3c3 100644 --- a/configs/h2rbox_v2/README.md +++ b/configs/h2rbox_v2/README.md @@ -45,7 +45,7 @@ HRSC ``` @inproceedings{yu2023h2rboxv2, title={H2RBox-v2: Incorporating Symmetry for Boosting Horizontal Box Supervised Oriented Object Detection}, - author={Yi Yu and Xue Yang and Qingyun Li and Yue Zhou and Gefan Zhang and Feipeng Da and Junchi Yan}, + author={Yi Yu and Xue Yang and Qingyun Li and Yue Zhou and and Feipeng Da and Junchi Yan}, year={2023}, booktitle={Advances in Neural Information Processing Systems} }