The paper introduces UNet 3+, a novel deep learning architecture for medical image segmentation, particularly focusing on organ segmentation. UNet 3+ leverages full-scale skip connections and deep supervision to enhance the accuracy and efficiency of the segmentation process. Full-scale skip connections integrate low-level details with high-level semantics from feature maps at different scales, while deep supervision learns hierarchical representations from aggregated feature maps. The method also includes a classification-guided module to reduce over-segmentation in non-organ images and a hybrid loss function to improve segmentation accuracy. Experimental results on liver and spleen datasets demonstrate that UNet 3+ outperforms existing methods, achieving higher Dice coefficients and more accurate boundary localization. The proposed approach is efficient, with fewer parameters, and shows promise for clinical applications in medical imaging.The paper introduces UNet 3+, a novel deep learning architecture for medical image segmentation, particularly focusing on organ segmentation. UNet 3+ leverages full-scale skip connections and deep supervision to enhance the accuracy and efficiency of the segmentation process. Full-scale skip connections integrate low-level details with high-level semantics from feature maps at different scales, while deep supervision learns hierarchical representations from aggregated feature maps. The method also includes a classification-guided module to reduce over-segmentation in non-organ images and a hybrid loss function to improve segmentation accuracy. Experimental results on liver and spleen datasets demonstrate that UNet 3+ outperforms existing methods, achieving higher Dice coefficients and more accurate boundary localization. The proposed approach is efficient, with fewer parameters, and shows promise for clinical applications in medical imaging.