What question did this study set out to answer?

To develop a lightweight underwater object detection model based on YOLOv8s that effectively handles image degradation, small targets, and computational resource constraints.

August 15, 2026Intelligent Marine Technology and SystemsOpen Access

TIBER-YOLO: an improved lightweight model for underwater object detection

View Full Paper

Ask AI

Bookmark

Authors

SASAMUEL ATTA ANTWI JAJoshua Yaw Amoako MEMichael Enyan

Discussion

Member takes

Overview

Model development study demonstrates improved detection accuracy with reduced computational complexity across underwater image datasets, suggesting enhanced feasibility for robotic edge deployment.

Key Points

To develop a lightweight underwater object detection model based on YOLOv8s that effectively handles image degradation, small targets, and computational resource constraints.
Integrated the TRACON module (triplet attention with receptive-field attention convolution) and Inner-WIoU loss function into the YOLOv8s framework.
Incorporated a bidirectional feature pyramid network (BiFPN) and an efficient multiscale partial convolution detector (EMPC-Detector) combining EMSConv and PConv.
Evaluated detection accuracy and computational efficiency across three benchmark datasets: DUO, UTDAC2020, and RUOD.
Achieved mAP@0.5 scores of 87.1% on DUO, 86.0% on UTDAC2020, and 86.1% on RUOD.
Reduced model size by 40.4%, total parameter count by 42.3%, and computational demand by 28.2% compared to baseline YOLOv8s.

Cite This Study

ANTWI et al. (2026) studied this question.

synapsesocial.com/papers/6a801a0e75c2e31742c86827 https://doi.org/10.1007/s44295-026-00111-9

View Full Paper

Ask AI

Bookmark

August 15, 2026Intelligent Marine Technology and SystemsOpen Access

TIBER-YOLO: an improved lightweight model for underwater object detection

View Full Paper

Ask AI

Bookmark

Authors

SASAMUEL ATTA ANTWI JAJoshua Yaw Amoako MEMichael Enyan

Discussion

Member takes

Overview

Model development study demonstrates improved detection accuracy with reduced computational complexity across underwater image datasets, suggesting enhanced feasibility for robotic edge deployment.

Key Points

To develop a lightweight underwater object detection model based on YOLOv8s that effectively handles image degradation, small targets, and computational resource constraints.
Integrated the TRACON module (triplet attention with receptive-field attention convolution) and Inner-WIoU loss function into the YOLOv8s framework.
Incorporated a bidirectional feature pyramid network (BiFPN) and an efficient multiscale partial convolution detector (EMPC-Detector) combining EMSConv and PConv.
Evaluated detection accuracy and computational efficiency across three benchmark datasets: DUO, UTDAC2020, and RUOD.
Achieved mAP@0.5 scores of 87.1% on DUO, 86.0% on UTDAC2020, and 86.1% on RUOD.
Reduced model size by 40.4%, total parameter count by 42.3%, and computational demand by 28.2% compared to baseline YOLOv8s.

Cite This Study

ANTWI et al. (2026) studied this question.

synapsesocial.com/papers/6a801a0e75c2e31742c86827 https://doi.org/10.1007/s44295-026-00111-9

View Full Paper

Ask AI

Bookmark