YOLO-LS: a novel deep learning framework for brain tumor segmentation in Magnetic Resonance Imaging
- 1The First Clinical Medical College, The Affiliated People's Hospital of Fujian University of Traditional Chinese Medicine, Fuzhou, Fujian, China
- 2The Affiliated People’s Hospital of Fujian University of Traditional Chinese Medicine
- 3The Affiliated People's Hospital of Fujian University of Traditional Chinese Medicine, Fuzhou, Fujian, China. [email protected]
Abstract
Brain tumors exhibit high heterogeneity in morphology, texture, and location, making accurate recognition and segmentation critical for clinical diagnosis, surgical planning, and prognosis evaluation. However, manual annotation of MRI scans is hindered by subjective bias and inefficiency. Furthermore, existing automated frameworks often face a trade-off between segmentation precision—particularly at infiltrative boundaries—and the computational efficiency required for deployment in resource-constrained environments. To address these challenges, this study proposes YOLO-LS (Lightweight Segmentation), an enhanced framework based on the YOLO11n-seg architecture designed for efficient detection and high-precision segmentation of brain tumors. The methodology introduces three key innovations: (1) integrating ShuffleNet V1 as a lightweight backbone to significantly reduce parameter count and computational complexity via pointwise grouped convolutions; (2) incorporating the DySample dynamic upsampling mechanism to mitigate the loss of fine-grained semantic details inherent in traditional interpolation, thereby improving the recovery of tumor boundaries; and (3) optimizing the neck network with a C3k2-PoolingFormer module to facilitate efficient cross-scale feature fusion and global context capture. The model was trained and tested on the Figshare dataset (3,064 images) using five-fold cross-validation and externally validated on an independent Kaggle dataset (300 images). Results demonstrate that YOLO-LS achieved a bounding box mAP50 of 0.953 ± 0.011, a Dice coefficient of 0.91 ± 0.01, and a 95% Hausdorff Distance (HD 95 ) of 4.35 ± 0.34 mm on the internal test set, the latter indicating superior boundary adherence compared to baseline models (paired t-test, p < 0.05 for all key metrics). Notably, the model reduced computational cost to 8.1 GFLOPs (a 15.6% reduction) while achieving an increase of 2.9% points in mAP50 compared to the baseline YOLO11n-seg. Comparative analyses with state-of-the-art architectures, including U-Net, SegNet, Swin-UNet, and VM-UNet, confirmed the efficacy of these improvements. Furthermore, Grad-CAM heatmaps validated the model’s precise focus on tumor core and edge regions. On the external test set, the model exhibited strong generalization with a Dice coefficient of 0.895. These findings indicate that YOLO-LS achieves an effective balance between precision, efficiency, and interpretability, demonstrating significant potential for assisting diagnostic workflows in resource-constrained clinical environments.
Related
- Hierarchical detection of neonatal brain pathologies using a two-stage deep learning framework from magnetic resonance imaging
- PANTHER Challenge Report: Cross-Domain Pancreatic Tumor Segmentation in Magnetic Resonance Imaging
- Automated optic chiasm segmentation in magnetic resonance imaging of multiple sclerosis patients using a multi-view, multi-class U-Net with limited annotated data
- Interpretable attention-deficit hyperactivity disorder diagnosis from functional magnetic resonance imaging via hybrid convolutional neural network–transformer learning and multi-level graph-based comprehensibility
- Functional magnetic resonance imaging in neuropathic pain: a bibliometric analysis incorporating machine learning
BibTeX
@article{chen2026yolols,
title = {{YOLO-LS: a novel deep learning framework for brain tumor segmentation in Magnetic Resonance Imaging}},
author = {Jinghui Chen and Yan Hu and Yang Tao and Zhipeng Sun and Lianxin Xie and Hongjia Zhao},
journal = {Scientific Reports},
year = {2026},
month = may,
doi = {10.1038/s41598-026-48825-4},
url = {https://doi.org/10.1038/s41598-026-48825-4},
note = {CC BY},
}
RIS
TY - JOUR
TI - YOLO-LS: a novel deep learning framework for brain tumor segmentation in Magnetic Resonance Imaging
AU - Jinghui Chen
AU - Yan Hu
AU - Yang Tao
AU - Zhipeng Sun
AU - Lianxin Xie
AU - Hongjia Zhao
PY - 2026
DA - 2026/05/13
JO - Scientific Reports
DO - 10.1038/s41598-026-48825-4
UR - https://doi.org/10.1038/s41598-026-48825-4
AB - Brain tumors exhibit high heterogeneity in morphology, texture, and location, making accurate recognition and segmentation critical for clinical diagnosis, surgical planning, and prognosis evaluation. However, manual annotation of MRI scans is hindered by subjective bias and inefficiency. Furthermore, existing automated frameworks often face a trade-off between segmentation precision—particularly at infiltrative boundaries—and the computational efficiency required for deployment in resource-constrained environments. To address these challenges, this study proposes YOLO-LS (Lightweight Segmentation), an enhanced framework based on the YOLO11n-seg architecture designed for efficient detection and high-precision segmentation of brain tumors. The methodology introduces three key innovations: (1) integrating ShuffleNet V1 as a lightweight backbone to significantly reduce parameter count and computational complexity via pointwise grouped convolutions; (2) incorporating the DySample dynamic upsampling mechanism to mitigate the loss of fine-grained semantic details inherent in traditional interpolation, thereby improving the recovery of tumor boundaries; and (3) optimizing the neck network with a C3k2-PoolingFormer module to facilitate efficient cross-scale feature fusion and global context capture. The model was trained and tested on the Figshare dataset (3,064 images) using five-fold cross-validation and externally validated on an independent Kaggle dataset (300 images). Results demonstrate that YOLO-LS achieved a bounding box mAP50 of 0.953 ± 0.011, a Dice coefficient of 0.91 ± 0.01, and a 95% Hausdorff Distance (HD 95 ) of 4.35 ± 0.34 mm on the internal test set, the latter indicating superior boundary adherence compared to baseline models (paired t-test, p < 0.05 for all key metrics). Notably, the model reduced computational cost to 8.1 GFLOPs (a 15.6% reduction) while achieving an increase of 2.9% points in mAP50 compared to the baseline YOLO11n-seg. Comparative analyses with state-of-the-art architectures, including U-Net, SegNet, Swin-UNet, and VM-UNet, confirmed the efficacy of these improvements. Furthermore, Grad-CAM heatmaps validated the model’s precise focus on tumor core and edge regions. On the external test set, the model exhibited strong generalization with a Dice coefficient of 0.895. These findings indicate that YOLO-LS achieves an effective balance between precision, efficiency, and interpretability, demonstrating significant potential for assisting diagnostic workflows in resource-constrained clinical environments.
SN - 2045-2322
ER -