This letter presents a novel segmentation approach that leverages dendritic neurons to tackle the challenges of medical imaging segmentation. The proposed method enhances segmentation accuracy based on a SegNet variant with an encoder-decoder structure, an upsampling index, and a deep supervision method. A dendritic neuron-based convolutional block is introduced to enable nonlinear feature mapping, improving the effectiveness of the approach. The method is evaluated on medical imaging segmentation datasets, showing superior performance compared to state-of-the-art methods. Medical image segmentation is crucial for accurate diagnosis and treatment planning. However, challenges arise due to the complexity and diversity of image content, making precise identification of regions of interest difficult. Existing feature mapping techniques in computer vision have limitations in capturing intricate details in medical images, affecting segmentation accuracy. Recent advances in deep learning have improved medical image segmentation, but methods like FCN and U-Net face challenges in capturing context and preserving details. SegNet addresses these issues through non-linear upsampling, but may suffer from information loss in non-uniform sampled images. To address these limitations, the authors propose a novel method called dendritic deep supervision neural network (DDNet), combining biologically interpretable dendritic neurons with deep supervision during training. This approach captures intricate details and fine-grained structures through nonlinear feature mapping, while deep supervision improves training efficiency and accuracy. The proposed method includes a deep supervision SegNet (DSegNet) module and a dendritic neuron model (DNM) module. DSegNet captures multi-scale information for accurate segmentation, while DNM performs nonlinear feature mapping. The DNM module uses a synapse layer, dendritic layer, membrane layer, and soma layer to process input features, enhancing the model's ability to capture intricate details in medical images. The method is evaluated on three datasets, showing superior performance in terms of segmentation accuracy. The loss function combines binary cross-entropy and focal loss, with parameters optimized for performance. The results demonstrate the effectiveness of the proposed approach in medical image segmentation.This letter presents a novel segmentation approach that leverages dendritic neurons to tackle the challenges of medical imaging segmentation. The proposed method enhances segmentation accuracy based on a SegNet variant with an encoder-decoder structure, an upsampling index, and a deep supervision method. A dendritic neuron-based convolutional block is introduced to enable nonlinear feature mapping, improving the effectiveness of the approach. The method is evaluated on medical imaging segmentation datasets, showing superior performance compared to state-of-the-art methods. Medical image segmentation is crucial for accurate diagnosis and treatment planning. However, challenges arise due to the complexity and diversity of image content, making precise identification of regions of interest difficult. Existing feature mapping techniques in computer vision have limitations in capturing intricate details in medical images, affecting segmentation accuracy. Recent advances in deep learning have improved medical image segmentation, but methods like FCN and U-Net face challenges in capturing context and preserving details. SegNet addresses these issues through non-linear upsampling, but may suffer from information loss in non-uniform sampled images. To address these limitations, the authors propose a novel method called dendritic deep supervision neural network (DDNet), combining biologically interpretable dendritic neurons with deep supervision during training. This approach captures intricate details and fine-grained structures through nonlinear feature mapping, while deep supervision improves training efficiency and accuracy. The proposed method includes a deep supervision SegNet (DSegNet) module and a dendritic neuron model (DNM) module. DSegNet captures multi-scale information for accurate segmentation, while DNM performs nonlinear feature mapping. The DNM module uses a synapse layer, dendritic layer, membrane layer, and soma layer to process input features, enhancing the model's ability to capture intricate details in medical images. The method is evaluated on three datasets, showing superior performance in terms of segmentation accuracy. The loss function combines binary cross-entropy and focal loss, with parameters optimized for performance. The results demonstrate the effectiveness of the proposed approach in medical image segmentation.