From 460d358c55f62f750bec813fb066f25a28303198 Mon Sep 17 00:00:00 2001 From: Joanne Lin Date: Sun, 22 Dec 2024 16:56:42 +0000 Subject: [PATCH] Updated ICASSP paper --- docs/_layouts/home.html | 6 +++-- docs/_posts/2024-02-05-LoIS.md | 31 ---------------------- docs/_posts/2024-12-20-LoIS.md | 29 ++++++++++++++++++++ docs/_site/2024/{02/05 => 12/20}/LoIS.html | 17 +++++------- 4 files changed, 40 insertions(+), 43 deletions(-) delete mode 100644 docs/_posts/2024-02-05-LoIS.md create mode 100644 docs/_posts/2024-12-20-LoIS.md rename docs/_site/2024/{02/05 => 12/20}/LoIS.html (93%) diff --git a/docs/_layouts/home.html b/docs/_layouts/home.html index 357ea69..0358759 100644 --- a/docs/_layouts/home.html +++ b/docs/_layouts/home.html @@ -93,7 +93,7 @@

Research Interests

News & Activities

@@ -130,7 +130,9 @@

Publications

Selected papers:
{% for post in site.posts %} {% if post.featured %} - {{ post.title }}

{{ post.authors | jsonify }}

+ + {{ post.title }} + {% for author in post.authors}{{ author }}, {% endfor %} {% endif %} {% endfor %} diff --git a/docs/_posts/2024-02-05-LoIS.md b/docs/_posts/2024-02-05-LoIS.md deleted file mode 100644 index d3211a9..0000000 --- a/docs/_posts/2024-02-05-LoIS.md +++ /dev/null @@ -1,31 +0,0 @@ ---- -title: Feature Denoising for Low-Light Instance Segmentation using Weighted Non-Local Blocks -alt_title: Low-Light Instance Segmentation -project: Low-Light Images and Videos -about: This project proposed a novel generic block for denoising the feature maps of low-light images, which can be easily integrated into existing instance segmentation frameworks. -authors: - names: [J. Lin, N. Anatrasirichai, D. Bull] - affiliations: [University of Bristol, University of Bristol, University of Bristol] - urls: [https://joannelin168.github.io/, https://pui-nantheera.github.io/, https://david-bull.github.io/] -funders: MyWorld Strength in Places -# collaborators: -# poster: -# poster_thumbnail: -paper: https://arxiv.org/abs/2402.18307 -arxiv: https://arxiv.org/abs/2402.18307 -supplementary: -github: -citation: | - @article{lin2024featdenoise, - title={Feature Denoising for Low-Light Instance Segmentation using Weighted Non-Local Blocks}, - author={Lin, Joanne and Anatrasirichai, Nantheera and Bull, David}, - year={2024}, - publisher={arXiv}} -featured: true - -layout: post ---- - -Instance segmentation for low-light imagery remains largely unexplored due to the challenges imposed by such conditions, for example shot noise due to low photon count, color distortions and reduced contrast. In this paper, we propose an end-to-end solution to address this challenging task. Based on Mask R-CNN, our proposed method implements weighted non-local (NL) blocks in the feature extractor. This integration enables an inherent denoising process at the feature level. As a result, our method eliminates the need for aligned ground truth images during training, thus supporting training on real-world low-light datasets. We introduce additional learnable weights at each layer in order to enhance the network's adaptability to real-world noise characteristics, which affect different feature scales in different ways. - -Experimental results show that the proposed method outperforms the pretrained Mask R-CNN with an Average Precision (AP) improvement of +10.0, with the introduction of weighted NL Blocks further enhancing AP by +1.0. diff --git a/docs/_posts/2024-12-20-LoIS.md b/docs/_posts/2024-12-20-LoIS.md new file mode 100644 index 0000000..a932856 --- /dev/null +++ b/docs/_posts/2024-12-20-LoIS.md @@ -0,0 +1,29 @@ +--- +title: Multi-Scale Denoising in the Feature Space for Low-Light Instance Segmentation +alt_title: Low-Light Instance Segmentation +project: Segmentation and Tracking in Low-Light +about: We propose a novel generic block for denoising the feature maps of low-light images, which can be easily integrated into existing instance segmentation frameworks. +authors: + names: [J. Lin, N. Anatrasirichai, D. Bull] + affiliations: [University of Bristol, University of Bristol, University of Bristol] + urls: [https://joannelin168.github.io/, https://pui-nantheera.github.io/, https://david-bull.github.io/] +funders: MyWorld Strength in Places +# collaborators: +# poster: +# poster_thumbnail: +paper: Currently unavailable +arxiv: https://arxiv.org/abs/2402.18307 +supplementary: +github: +citation: | + @article{lin2024lowlightsegm, + title={Multi-Scale Denoising in the Feature Space for Low-Light Instance Segmentation}, + author={Lin, Joanne and Anatrasirichai, Nantheera and Bull, David}, + year={2025}, + publisher={ICASSP}} +featured: true + +layout: post +--- + +Instance segmentation for low-light imagery remains largely unexplored due to the challenges imposed by such conditions, for example shot noise due to low photon count, color distortions and reduced contrast. In this paper, we propose an end-to-end solution to address this challenging task. Our proposed method implements weighted non-local blocks (wNLB) in the feature extractor. This integration enables an inherent denoising process at the feature level. As a result, our method eliminates the need for aligned ground truth images during training, thus supporting training on real-world low-light datasets. We introduce additional learnable weights at each layer in order to enhance the network's adaptability to real-world noise characteristics, which affect different feature scales in different ways. Experimental results on several object detectors show that the proposed method outperforms the pretrained networks with an Average Precision (AP) improvement of at least +7.6, with the introduction of wNLB further enhancing AP by upto +1.3. \ No newline at end of file diff --git a/docs/_site/2024/02/05/LoIS.html b/docs/_site/2024/12/20/LoIS.html similarity index 93% rename from docs/_site/2024/02/05/LoIS.html rename to docs/_site/2024/12/20/LoIS.html index f2be87c..ccbe223 100644 --- a/docs/_site/2024/02/05/LoIS.html +++ b/docs/_site/2024/12/20/LoIS.html @@ -106,23 +106,20 @@

Downloads

-
+

Citation

- 
@article{hill2020habnet,
-    title={HABNet: Machine Learning, Remote Sensing-Based Detection of Harmful Algal Blooms},
-    author={Hill, Paul R and Kumar, Anurag and Temimi, Marouane and Bull, David R},
-    journal={IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing},
-    volume={13},
-    pages={3229--3239},
-    year={2020},
-    publisher={IEEE}}
+            
@article{lin2024lowlightsegm,
+title={Multi-Scale Denoising in the Feature Space for Low-Light Instance Segmentation},
+author={Lin, Joanne and Anatrasirichai, Nantheera and Bull, David},
+year={2025},
+publisher={ICASSP}}