2026-10-11T04:19:59Z https://api.figshare.com/v2/oai

oai:figshare.com:article/33814453 2026-09-15T17:58:21Z category_146 category_15 category_39 category_734 category_931 portal_5 item_type_3 month_year_09_2026

<p>Performance across five independent runs (mAP@50).</p> Mahdi Hamzaoui (24950158) Leila Bousbia (24950161) Mohamed Ould-Elhassen Aoueileyine (19743433) Imen Filali (23739873) Ridha Bouallegue (19769292) Space Science Neuroscience Ecology Biological Sciences not elsewhere classified Information Systems not elsewhere classified traditional recognition approaches partially occluded fish overall production efficiency fish species remains experimental results indicate enhanced architecture designed counter low contrast underwater fish recognition baseline yolov11 model world aquaculture systems clear performance boost improvements highlight yolo cbnet &# 8217 enhanced detection framework yolov11 framework underwater detection xlink "> uneven lighting suspended particles stage strategy significantly limit robust attention reliable solution reduced visibility pivotal challenge method adopts health monitoring dependable monitoring combination intended color distortion bifpn ), achieved 0 accurate identification <p>Performance across five independent runs (mAP@50).</p> 2026-09-15T17:57:59Z Dataset Dataset 10.1371/journal.pone.0341525.t006 https://figshare.com/articles/dataset/_p_Performance_across_five_independent_runs_mAP_50_p_/33814453 CC BY 4.0