From image-level to pixel-level labeling: A weakly-supervised learning method for identifying aquaculture ponds using iterative anti-adversarial attacks guided by aquaculture features DOI Creative Commons
Boyi Li, Adu Gong, Jiaming Zhang

et al.

International Journal of Applied Earth Observation and Geoinformation, Journal Year: 2024, Volume and Issue: 132, P. 104023 - 104023

Published: July 15, 2024

Aquaculture mapping is essential for monitoring and managing aquaculture resources. However, accurately geotargeting individual ponds from medium-resolution remote sensing imagery remains challenging, convolutional deep learning methods identifying require labor-intensive pixel-level annotations. This paper presents a novel weakly-supervised method to derive labels image-level annotations ponds. Our approach uses iterative anti-adversarial attacks refine localization results multi-scale class activation maps (CAMs). The improved integrates two regularization guided by features form joint loss function adversarial samples: discriminative water region suppression non-aquaculture suppression. We also propose an feature termed CFNDWI constrain the generate high-quality pseudo-labels. As result, pseudo-labels are used train semantic segmentation networks evaluated performance of our using commonly-used backbones on 10 m Sentinel-2 imagery. achieves Intersection over Union (IoU) values 0.618–0.655 pseudo-label generation, IoU 0.664–0.708 segmentation, outperforming state-of-the-art public datasets. effectiveness each module was testified through ablation experiments. leverages knowledge-driven guide data-driven process, addressing lack datasets model training. code implementing will be accessible at https://github.com/designer1024/WSLM-AQ.

Language: Английский

Phytoplankton composition and metabolomic profiles in aquaculture systems: A case study in Brazil's natural lakes DOI
Éryka Costa de Almeida, Fernanda Rios Jacinavicius, Larissa Souza Passos

et al.

Aquaculture, Journal Year: 2025, Volume and Issue: unknown, P. 742135 - 742135

Published: Jan. 1, 2025

Language: Английский

Citations

0
In-water Bacillus species probiotic improved water quality, growth, hemato-biochemical profile, immune regulatory genes and resistance of Nile tilapia to Aspergillus flavus infection DOI
El‐Sayed Hemdan Eissa, Ekemini Moses Okon, Abdel‐Wahab A. Abdel‐Warith

et al.

Aquaculture International, Journal Year: 2024, Volume and Issue: 32(6), P. 7087 - 7102

Published: April 23, 2024

Language: Английский

Citations

3
Cyanobacteria: role in sustainable biomanufacturing and nitrogen fixation DOI
Taufiq Nawaz, Shah Fahad, Liping Gu

et al.

Biofuels Bioproducts and Biorefining, Journal Year: 2024, Volume and Issue: unknown

Published: Sept. 19, 2024

Abstract Cyanobacteria, renowned for their nitrogen‐fixing characteristics, are important sustainable biomanufacturing and agricultural innovation. This review explores the synergy between cyanobacteria nitrogen fixation, highlighting potential to revolutionize biobased compound production reduce ecological impact of traditional sources. It focuses on genetic enhancements synthetic biology techniques, which transform these microorganisms into providers. Current applications range from enhancement cutting‐edge biotechnology, consequences cyanobacterial fixation. Challenges persist, however, requiring a meticulous analysis ecological, regulatory, scalability concerns. The untapped in fixation promises significant shift environmental stewardship. aim this article is inspire high‐impact research transformative biotechnology sustainability.

Language: Английский

Citations

3
Hepatocyte apoptosis is triggered by hepatic inflammation in common carp acutely exposed to microcystin-LR or chronically exposed to Microcystis DOI Creative Commons
Haoyang Zhao,

Kehui Sun,

X. Nan

et al.

Ecotoxicology and Environmental Safety, Journal Year: 2024, Volume and Issue: 286, P. 117230 - 117230

Published: Oct. 22, 2024

Language: Английский

Citations

3
Comparative Analysis of Cyanotoxins in Fishponds in Nigeria and South Africa DOI Creative Commons

Odo J. Bassey,

Jabulani R. Gumbo, Munyaradzi Mujuru

et al.

Microbiology Research, Journal Year: 2024, Volume and Issue: 15(2), P. 447 - 456

Published: March 24, 2024

Over the decades, aquaculture sector has witnessed substantial growth, contributing significantly to nation’s economy. However, menace of CyanoHABs threatens sustainability fish farming. Considering possible hazards linked cyanotoxins in food and water, a comparative study design between commercial Nigeria South Africa was employed investigate water from fishponds. Six fishponds Calabar Municipality—Nigeria Duthuni—South with varying climatic zones were selected. Water samples ponds collected at intervals during different seasons (summer, winter, dry, wet seasons) capture climate-induced variation. Liquid chromatography–mass spectrometry (LCMS) combination metabolites database used for identification toxic cyanometabolites samples. The molecular networking approach, coupled Global Natural Products Social Molecular Networking (GNPS) CANOPUS annotation, enabled putative cyanometabolites. resulting network unveiled discernible clusters representing related molecule families, aiding both known unfamiliar analogues. Furthermore, revealed that shared specific metabolites, including ethanesulfonic acid, pheophorbide A, cholic phenylalanine, amyl amine, phosphocholine (PC), sulfonic despite variations location, local factors, sampling sites. showed presence multiple cyanotoxin classes wet, summer water. Aflatoxin identified all sites (N1, N2, N3). Duthuni, Africa, (P1, P2, P3) exhibited microginins microcystins. All displayed widespread occurrence anabaenopeptins, aplysiatoxins, aflatoxin, microcolins, marabmids selected summer. In conclusion, untargeted metabolome analysis, guided by GNPS, proved highly effective identifying non-toxic

Language: Английский

Citations

1
From image-level to pixel-level labeling: A weakly-supervised learning method for identifying aquaculture ponds using iterative anti-adversarial attacks guided by aquaculture features DOI Creative Commons
Boyi Li, Adu Gong, Jiaming Zhang

et al.

International Journal of Applied Earth Observation and Geoinformation, Journal Year: 2024, Volume and Issue: 132, P. 104023 - 104023

Published: July 15, 2024

Aquaculture mapping is essential for monitoring and managing aquaculture resources. However, accurately geotargeting individual ponds from medium-resolution remote sensing imagery remains challenging, convolutional deep learning methods identifying require labor-intensive pixel-level annotations. This paper presents a novel weakly-supervised method to derive labels image-level annotations ponds. Our approach uses iterative anti-adversarial attacks refine localization results multi-scale class activation maps (CAMs). The improved integrates two regularization guided by features form joint loss function adversarial samples: discriminative water region suppression non-aquaculture suppression. We also propose an feature termed CFNDWI constrain the generate high-quality pseudo-labels. As result, pseudo-labels are used train semantic segmentation networks evaluated performance of our using commonly-used backbones on 10 m Sentinel-2 imagery. achieves Intersection over Union (IoU) values 0.618–0.655 pseudo-label generation, IoU 0.664–0.708 segmentation, outperforming state-of-the-art public datasets. effectiveness each module was testified through ablation experiments. leverages knowledge-driven guide data-driven process, addressing lack datasets model training. code implementing will be accessible at https://github.com/designer1024/WSLM-AQ.

Language: Английский

Citations

1