From 007d13ad02197802dd4481d8d74a7136d9b80286 Mon Sep 17 00:00:00 2001 From: Ehtesham Iqbal <79190474+Ehteshamciitwah@users.noreply.github.com> Date: Mon, 24 Oct 2022 10:48:48 +0900 Subject: [PATCH 1/8] Update README.md --- README.md | 14 +------------- 1 file changed, 1 insertion(+), 13 deletions(-) diff --git a/README.md b/README.md index cbb83ef..9f2becd 100644 --- a/README.md +++ b/README.md @@ -2,11 +2,6 @@ # MSANet: Multi-Similarity and Attention Guidance for Boosting Few-Shot Segmentation This is the official implementation of the paper [MSANet: Multi-Similarity and Attention Guidance for Boosting Few-Shot Segmentation](https://arxiv.org/abs/2206.09667) -[![PWC](https://img.shields.io/endpoint.svg?url=https://paperswithcode.com/badge/msanet-multi-similarity-and-attention-1/few-shot-semantic-segmentation-on-coco-20i-1)](https://paperswithcode.com/sota/few-shot-semantic-segmentation-on-coco-20i-1?p=msanet-multi-similarity-and-attention-1) -[![PWC](https://img.shields.io/endpoint.svg?url=https://paperswithcode.com/badge/msanet-multi-similarity-and-attention-1/few-shot-semantic-segmentation-on-coco-20i-5)](https://paperswithcode.com/sota/few-shot-semantic-segmentation-on-coco-20i-5?p=msanet-multi-similarity-and-attention-1) -[![PWC](https://img.shields.io/endpoint.svg?url=https://paperswithcode.com/badge/msanet-multi-similarity-and-attention-1/few-shot-semantic-segmentation-on-pascal-5i-1)](https://paperswithcode.com/sota/few-shot-semantic-segmentation-on-pascal-5i-1?p=msanet-multi-similarity-and-attention-1) -[![PWC](https://img.shields.io/endpoint.svg?url=https://paperswithcode.com/badge/msanet-multi-similarity-and-attention-1/few-shot-semantic-segmentation-on-pascal-5i-5)](https://paperswithcode.com/sota/few-shot-semantic-segmentation-on-pascal-5i-5?p=msanet-multi-similarity-and-attention-1) - Authors: Ehtesham Iqbal, Sirojbek Safarov, Seongdeok Bang > **Abstract:** *Few-shot segmentation aims to segment unseen-class objects given only a handful of densely labeled samples. Prototype learning, where the support feature yields a singleor several prototypes by averaging global and local object information, has been widely used in FSS. However, utilizing only prototype vectors may be insufficient to represent the features for all training data. To extract abundant features and make more precise predictions, we propose a Multi-Similarity and Attention Network (MSANet) including two novel modules, a multi-similarity module and an attention module. The multi-similarity module exploits multiple feature-maps of support images and query images to estimate accurate semantic relationships. The attention module instructs the network to concentrate on class-relevant information. The network is tested on standard FSS datasets, PASCAL-5i 1-shot, PASCAL-5i 5-shot, COCO-20i 1-shot, and COCO-20i 5-shot. The MSANet with the backbone of ResNet-101 achieves the state-of-the-art performance for all 4-benchmark datasets with mean intersection over union (mIoU) of 69.13%, 73.99%, 51.09%, 56.80%, respectively.* @@ -78,12 +73,5 @@ Performance comparison with the state-of-the-art approaches (*i.e.*, [HSNet](htt This repo is mainly built based on [PFENet](https://github.com/dvlab-research/PFENet), [HSNet](https://github.com/juhongm999/hsnet), and [BAM](https://github.com/chunbolang/BAM). Thanks for their great work! ### BibTeX -If you find this research useful, please consider citing: -````BibTeX -@article{MSANet2022, - title={MSANet: Multi-Similarity and Attention Guidance for Boosting Few-Shot Segmentation}, - author={Ehtesham Iqbal, Sirojbek Safarov, Seongdeok Bang}, - journal={arXiv preprint arXiv:2206.09667}, - year={2022} -} + ```` From 98e97fb37e6a37a265958c81b049baa0188c6c16 Mon Sep 17 00:00:00 2001 From: Ehtesham Iqbal <79190474+Ehteshamciitwah@users.noreply.github.com> Date: Mon, 24 Oct 2022 10:49:15 +0900 Subject: [PATCH 2/8] Update README.md --- README.md | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/README.md b/README.md index 9f2becd..0d9d192 100644 --- a/README.md +++ b/README.md @@ -1,6 +1,6 @@ # MSANet: Multi-Similarity and Attention Guidance for Boosting Few-Shot Segmentation -This is the official implementation of the paper [MSANet: Multi-Similarity and Attention Guidance for Boosting Few-Shot Segmentation](https://arxiv.org/abs/2206.09667) +This is the official implementation of the paper [MSANet: Multi-Similarity and Attention Guidance for Boosting Few-Shot Segmentation] Authors: Ehtesham Iqbal, Sirojbek Safarov, Seongdeok Bang From cd69adf0da248c64fe369aed4791ea17921bf17a Mon Sep 17 00:00:00 2001 From: Ehtesham Iqbal <79190474+Ehteshamciitwah@users.noreply.github.com> Date: Mon, 24 Oct 2022 10:51:49 +0900 Subject: [PATCH 3/8] Update README.md --- README.md | 3 --- 1 file changed, 3 deletions(-) diff --git a/README.md b/README.md index 0d9d192..7077713 100644 --- a/README.md +++ b/README.md @@ -1,10 +1,7 @@ # MSANet: Multi-Similarity and Attention Guidance for Boosting Few-Shot Segmentation -This is the official implementation of the paper [MSANet: Multi-Similarity and Attention Guidance for Boosting Few-Shot Segmentation] -Authors: Ehtesham Iqbal, Sirojbek Safarov, Seongdeok Bang -> **Abstract:** *Few-shot segmentation aims to segment unseen-class objects given only a handful of densely labeled samples. Prototype learning, where the support feature yields a singleor several prototypes by averaging global and local object information, has been widely used in FSS. However, utilizing only prototype vectors may be insufficient to represent the features for all training data. To extract abundant features and make more precise predictions, we propose a Multi-Similarity and Attention Network (MSANet) including two novel modules, a multi-similarity module and an attention module. The multi-similarity module exploits multiple feature-maps of support images and query images to estimate accurate semantic relationships. The attention module instructs the network to concentrate on class-relevant information. The network is tested on standard FSS datasets, PASCAL-5i 1-shot, PASCAL-5i 5-shot, COCO-20i 1-shot, and COCO-20i 5-shot. The MSANet with the backbone of ResNet-101 achieves the state-of-the-art performance for all 4-benchmark datasets with mean intersection over union (mIoU) of 69.13%, 73.99%, 51.09%, 56.80%, respectively.*

From 31322fa709a2716c65479beb0959eb60141c3f77 Mon Sep 17 00:00:00 2001 From: Ehtesham Iqbal <79190474+Ehteshamciitwah@users.noreply.github.com> Date: Mon, 24 Oct 2022 11:27:56 +0900 Subject: [PATCH 4/8] Update README.md --- README.md | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/README.md b/README.md index 7077713..f2182b8 100644 --- a/README.md +++ b/README.md @@ -21,7 +21,7 @@ - COCO-20i: [COCO2014](https://cocodataset.org/#download) - Download the [data](https://aivkr-my.sharepoint.com/:f:/g/personal/safarov_sirojbek_aiv_ai/EsTvSTPyp_NCq-RIifEAnSMBy8BfNX2iVlfZZ0nSnwi3RQ?e=d3OWUj) lists (.txt files) and put them into the `MSANet/lists` directory. + Download the [data](https://drive.google.com/drive/folders/1GdfpfkBEVNnOENYepPxjnbIHEMavIIOw?usp=sharing) lists (.txt files) and put them into the `MSANet/lists` directory. ### Models From f5f575d75abd2c5da8fcac9799454e9b5b13f3f6 Mon Sep 17 00:00:00 2001 From: Ehtesham Iqbal <79190474+Ehteshamciitwah@users.noreply.github.com> Date: Mon, 24 Oct 2022 11:43:27 +0900 Subject: [PATCH 5/8] Update README.md --- README.md | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/README.md b/README.md index f2182b8..cf33e44 100644 --- a/README.md +++ b/README.md @@ -21,7 +21,7 @@ - COCO-20i: [COCO2014](https://cocodataset.org/#download) - Download the [data](https://drive.google.com/drive/folders/1GdfpfkBEVNnOENYepPxjnbIHEMavIIOw?usp=sharing) lists (.txt files) and put them into the `MSANet/lists` directory. + Download the [data](https://file.io/0mbdYoWJLI1f) lists (.txt files) and put them into the `MSANet/lists` directory. ### Models From 455b8d61fd98be52b510b5ccf044061c1b76ca9e Mon Sep 17 00:00:00 2001 From: Ehtesham Iqbal <79190474+Ehteshamciitwah@users.noreply.github.com> Date: Mon, 24 Oct 2022 12:08:33 +0900 Subject: [PATCH 6/8] Update README.md --- README.md | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/README.md b/README.md index cf33e44..6a72b0d 100644 --- a/README.md +++ b/README.md @@ -21,7 +21,7 @@ - COCO-20i: [COCO2014](https://cocodataset.org/#download) - Download the [data](https://file.io/0mbdYoWJLI1f) lists (.txt files) and put them into the `MSANet/lists` directory. + Download the [data](https://drive.google.com/uc?export=download&id=1J1vPgctx8ojlkswMZS9Ccc8_bU6Zuej4) and put them into the `MSANet/lists` directory. ### Models From 122b6887bb381da17419beccd55f5ffc982ba6c3 Mon Sep 17 00:00:00 2001 From: Ehtesham Iqbal <79190474+Ehteshamciitwah@users.noreply.github.com> Date: Mon, 24 Oct 2022 12:24:43 +0900 Subject: [PATCH 7/8] Update README.md --- README.md | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) diff --git a/README.md b/README.md index 6a72b0d..ed2e75c 100644 --- a/README.md +++ b/README.md @@ -25,9 +25,9 @@ ### Models -- Download the pre-trained backbones from [here](https://aivkr-my.sharepoint.com/:f:/g/personal/safarov_sirojbek_aiv_ai/EnGqMXVD5N5HrNgAKDpx0kUB0xo720V5L0VWRwHvVOKukw?e=90JVzl) and put them into the `MSANet/initmodel` directory. -- Download our trained base learners from [OneDrive](https://aivkr-my.sharepoint.com/:f:/g/personal/safarov_sirojbek_aiv_ai/EsAKfmsEqp5DmJ4gaiUtRqUB9b256ObgzfVZ-U-R50IlFw?e=z5HIM6) and put them under `initmodel/PSPNet`. -- We provide all trained MSANet [models](https://aivkr-my.sharepoint.com/:f:/g/personal/safarov_sirojbek_aiv_ai/EjDn3jyTVWFHso3uX8_AgSgBj1y_nB3hQ0wP8RS9aE6Cdw?e=DbT3eH) for performance evaluation. _Backbone: VGG16 & ResNet50; Dataset: PASCAL-5i & COCO-20i; Setting: 1-shot & 5-shot_. +- Download the pre-trained backbones from [here](https://drive.google.com/uc?export=download&id=1bWVt8OZt2pSDtXn_XLDbR5obSTVrUpN9) and put them into the `MSANet/initmodel` directory. +- Download our trained base learners from [OneDrive](https://drive.google.com/uc?export=download&id=1pPOC2rsSPMTm7B4Dr2b1yJnsMqa20u98) and put them under `initmodel/PSPNet`. +- We provide all trained MSANet [models](https://drive.google.com/uc?export=download&id=14ILPhyKXva9N8pZB495T5DC5081v9IJj) for performance evaluation. _Backbone: VGG16 & ResNet50; Dataset: PASCAL-5i & COCO-20i; Setting: 1-shot & 5-shot_. ### Scripts From b6156b2d03d282fb876cceacece542444b464fbe Mon Sep 17 00:00:00 2001 From: Ehtesham Iqbal <79190474+Ehteshamciitwah@users.noreply.github.com> Date: Mon, 24 Oct 2022 12:26:52 +0900 Subject: [PATCH 8/8] Update README.md --- README.md | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) diff --git a/README.md b/README.md index ed2e75c..b79cd04 100644 --- a/README.md +++ b/README.md @@ -26,7 +26,7 @@ ### Models - Download the pre-trained backbones from [here](https://drive.google.com/uc?export=download&id=1bWVt8OZt2pSDtXn_XLDbR5obSTVrUpN9) and put them into the `MSANet/initmodel` directory. -- Download our trained base learners from [OneDrive](https://drive.google.com/uc?export=download&id=1pPOC2rsSPMTm7B4Dr2b1yJnsMqa20u98) and put them under `initmodel/PSPNet`. +- Download our trained base learners from [Drive](https://drive.google.com/uc?export=download&id=1pPOC2rsSPMTm7B4Dr2b1yJnsMqa20u98) and put them under `initmodel/PSPNet`. - We provide all trained MSANet [models](https://drive.google.com/uc?export=download&id=14ILPhyKXva9N8pZB495T5DC5081v9IJj) for performance evaluation. _Backbone: VGG16 & ResNet50; Dataset: PASCAL-5i & COCO-20i; Setting: 1-shot & 5-shot_. ### Scripts @@ -35,7 +35,7 @@ ### Performance -Performance comparison with the state-of-the-art approaches (*i.e.*, [HSNet](https://github.com/juhongm999/hsnet), [BAM](https://github.com/chunbolang/BAM) and [VAT](https://github.com/Seokju-Cho/Volumetric-Aggregation-Transformer) in terms of **average** **mIoU** across all folds. +Performance comparison with the state-of-the-art approaches (*i.e.*, [HSNet], [BAM] and [VAT] in terms of **average** **mIoU** across all folds. 1. ##### PASCAL-5i @@ -67,7 +67,7 @@ Performance comparison with the state-of-the-art approaches (*i.e.*, [HSNet](htt ## References -This repo is mainly built based on [PFENet](https://github.com/dvlab-research/PFENet), [HSNet](https://github.com/juhongm999/hsnet), and [BAM](https://github.com/chunbolang/BAM). Thanks for their great work! + ### BibTeX