Fig. 6: Heatmaps of feature extraction in various models. | Communications Engineering

Fig. 6: Heatmaps of feature extraction in various models.

From: A deep learning framework based on structured space model for detecting small objects in complex underwater environments

Fig. 6

a Performance of different models on the DUO datasets. b Performance of different models on the URPC2021 datasets. c Visualization of heatmap effects during feature extraction using the GradCAM method. The left side shows the original input image, followed by the visualization results of the baseline, YOLOv10, Mamba-YOLO, and UWNet models.

Back to article page