🤖 AI Summary
Medical image segmentation faces challenges in deployment on mobile healthcare devices due to computational constraints and poor cross-dataset robustness of lightweight models. To address these issues, this paper proposes a lightweight, plug-and-play segmentation framework. Methodologically, it introduces (1) a reparameterizable fusible module that decouples training and inference while preserving vanilla CNN architecture and enhancing generalization; and (2) an enhanced deep supervision strategy integrated with a MobileNetV3-Large backbone, coupled with structural simplification during inference for acceleration. Evaluated on five medical datasets—including ISIC 2016 and BUSI—the framework achieves significant reductions in model parameters and FLOPs, attaining optimal accuracy-efficiency trade-offs. It enables real-time, edge-side segmentation while meeting clinical deployment requirements.
📝 Abstract
While large models have achieved significant progress in computer vision, challenges such as optimization complexity, the intricacy of transformer architectures, computational constraints, and practical application demands highlight the importance of simpler model designs in medical image segmentation. This need is particularly pronounced in mobile medical devices, which require lightweight, deployable models with real-time performance. However, existing lightweight models often suffer from poor robustness across datasets, limiting their widespread adoption. To address these challenges, this paper introduces LV-UNet, a lightweight and vanilla model that leverages pre-trained MobileNetv3-Large backbones and incorporates fusible modules. LV-UNet employs an enhanced deep training strategy and switches to a deployment mode during inference by re-parametrization, significantly reducing parameter count and computational overhead. Experimental results on ISIC 2016, BUSI, CVC-ClinicDB, CVC-ColonDB, and Kvair-SEG datasets demonstrate a better trade-off between performance and the computational load. The code will be released at https://github.com/juntaoJianggavin/LV-UNet.