Correction: Xu et al. MC-ASFF-ShipYOLO: Improved Algorithm for Small-Target and Multi-Scale Ship Detection for Synthetic Aperture Radar (SAR) Images. Sensors 2025, 25, 2940

Abstract

Full text

Legends of Figures 1 and 3

In the original publication [1], the source attribution for Figures 1 and 3 were inadvertently omitted. The corrected Figures 1 and 3 appears below.

Figure 1. (a) Ship targets in complex inland river environments; (b) scene showing densely distributed small ship targets; (c) examples of multi-scale ship targets after zero-padding processing of the SSDD dataset. Note that (a,b) are from the HRSID dataset. Red bounding boxes highlight real ships.Figure 3. Overall framework of the MC-ASFF-ShipYOLO model based on YOLO11 improvements, with improved modules highlighted in red boxes. The foundational YOLO11 network structure is independently drawn based on the Ultralytics configuration files [43].

Missing Citation

In the original publication, ref. [43] was not cited. The citation has now been inserted in Section 1, Paragraph 6 and should read as follows:

To overcome these limitations, we propose MC-ASFF-ShipYOLO (Monte Carlo Attention—Adaptively Spatial Feature Fusion—ShipYOLO), an improved single-stage detector based on YOLO11 [43,44]. Our main contributions are as follows.

In addition, the citation has been inserted in Section 2.4, Paragraph 1 and should read as follows:

MC-ASFF-ShipYOLO is an improved model based on the YOLO11 object detection network. As a YOLO series model released on 30 September 2024 [43], YOLO11 achieves significant improvements in feature extraction capability and small object detection performance compared to previous versions, with the advantages of higher precision, fewer parameters, and faster inference speed.

The citation has also been inserted in Section 3.1, Paragraph 1 and should read as follows:

In this section, we chose the two-stage detectors, alongside the high-efficiency neural network model EfficientNet, and incorporated various editions of the YOLO series (such as YOLOv8 [52], YOLOv9 [53], YOLOv10 [54], YOLO11 [43], and YOLOv12 [55]) to conduct comparative experiments. Table 3 provides a detailed comparison of the experimental outcomes.

In the original publication, ref. [45] was not cited. The citation has now been inserted in Section 2.1, Paragraph 1 and should read as follows:

This study employs a hybrid dataset, constructed from HRSID and SSDD, as the basis for model training and performance evaluation. Details of the hybrid dataset can be found in the Supplementary Materials. Released in 2020, HRSID [45] is a high-resolution SAR image dataset designed for ship detection and segmentation tasks.

In the original publication, ref. [46] was not cited. The citation has now been inserted in Table 2, as shown below:ModelsBackboneBatch Sizelr0lrf****NMSFaster R-CNNR50160.010.00010.6R101120.010.00010.7Cascade R-CNNR50160.010.00010.6R101160.010.00010.7EfficientNet [46]Eff-b360.010.0001-

In the original publication, refs. [52–55] were not cited. The citation [52] has now been inserted in Section 3, Paragraph 1 and should read as follows:

This section provides a detailed evaluation of the MC-ASFF-ShipYOLO model’s performance on the constructed hybrid SAR ship dataset and systematically compares it with classical object detection networks. We conduct a comprehensive assessment of the improved model’s effectiveness in SAR ship detection tasks through both quantitative metrics and qualitative analysis. YOLO11 is an improved version of YOLOv8 [52].

The citations [52–55] have now been inserted in Section 3.1, Paragraph 1 and should read:

In this section, we chose the two-stage detectors, alongside the high-efficiency neural network model EfficientNet, and incorporated various editions of the YOLO series (such as YOLOv8 [52], YOLOv9 [53], YOLOv10 [54], YOLO11 [43], and YOLOv12 [55]) to conduct comparative experiments. Table 3 provides a detailed comparison of the experimental outcomes.

Text Correction

In the original publication, the model was referred to using non-standard terminology. The name “YOLO12” has been updated to the standardized designation “YOLOv12” for technical accuracy. A correction has been made to Results, Section 3.1, Paragraph 1:

In this section, we chose the two-stage detectors, alongside the high-efficiency neural network model EfficientNet, and incorporated various editions of the YOLO series (such as YOLOv8 [52], YOLOv9 [53], YOLOv10 [54], YOLO11 [43], and YOLOv12 [55]) to conduct comparative experiments. Table 3 provides a detailed comparison of the experimental outcomes.

A correction has also been made to Table 3: ModelsBackboneor Size $[eqn]$ /% $[eqn]$ /% $[eqn]$ /% $[eqn]$ **/%****Params (M)****FPS (img/s)**YOLOv8n90.8582.8390.6764.053.01666.67s90.7484.7591.8165.2411.13370.37YOLOv9t91.1183.8291.5565.761.97555.57s90.9084.9491.9767.537.17333.33YOLOv10n88.9081.5590.1165.202.70666.67s91.0983.4091.7666.778.04384.62YOLO11n90.4181.5689.2063.922.58555.57s92.4884.2192.7867.409.41370.37YOLOv12n90.9381.0890.4366.042.56416.67s91.3982.2191.5966.249.23243.90EfficientNetEff-b374.9189.2988.0061.3018.34306.30Faster R-CNNR5084.1882.8783.4060.8041.35308.10R10184.1883.0883.2060.5060.34324.00Cascade R-CNNR5084.5183.9884.5063.1069.15317.50R10184.9784.1383.8062.8088.14317.10MS-ASFF-ShipYOLO (Ours)-93.8786.8494.5670.4460.28232.56

Supplementary Materials

A direct mention of the Supplementary Materials has been added to Section 2.1, Paragraph 1 and should read as follows:

This study employs a hybrid dataset, constructed from HRSID and SSDD, as the basis for model training and performance evaluation. Details of the hybrid dataset can be found in the Supplementary Materials. Released in 2020, HRSID [45] is a high-resolution SAR image dataset designed for ship detection and segmentation tasks.

Supplementary Materials: The raw data supporting the findings of this study are available at the following link: https://pan.baidu.com/s/15XwGxaZddf94N_V_MkCqgw?pwd=kzpk (accessed on 30 December 2025).