مقایسهٔ روش‌ها

	بخش‌بندی نمونه تطبیقی دامنه	یادگیری انتقالی با بخش‌بندی نمونه (Transfer Learning with Instance Segmentation)
حوزه	یادگیری عمیق	یادگیری عمیق
خانواده	Machine learning	Machine learning
سال پیدایش	2018–2021	2017 (Mask R-CNN); transfer learning paradigm: 2010
پدیدآور	Chen, Y. et al. (domain-adaptive detection); extended to instance segmentation by multiple groups ~2019–2021	He, K. et al. (Mask R-CNN); transfer learning framework: Pan & Yang
نوع	Domain adaptation + instance segmentation	Transfer learning applied to instance segmentation
منبع بنیادین	Chen, Y., Li, W., Sakaridis, C., Dai, D., & Van Gool, L. (2018). Domain Adaptive Faster RCNN for Object Detection in the Wild. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 3339–3348.	He, K., Gkioxari, G., Dollar, P., & Girshick, R. (2017). Mask R-CNN. Proceedings of the IEEE International Conference on Computer Vision (ICCV), 2961–2969.
نام‌های دیگر	DA-InstanceSeg, cross-domain instance segmentation, domain adaptation for instance segmentation, unsupervised domain adaptive Mask R-CNN	pretrained instance segmentation, fine-tuned Mask R-CNN, transfer learning for panoptic segmentation, domain-adapted instance segmentation
مرتبط	3	4
خلاصه	Domain-adaptive instance segmentation extends Mask R-CNN-style architectures to operate across distribution shifts — training on a labeled source domain (e.g., synthetic renderings or daytime images) and adapting to an unlabeled or weakly labeled target domain (e.g., real scenes or nighttime footage). Adversarial feature alignment and self-training close the domain gap at both image-level and instance-level granularity.	Transfer learning with instance segmentation reuses a backbone convolutional network pretrained on a large image corpus (typically ImageNet or COCO) as the feature extractor for an instance segmentation model such as Mask R-CNN, then fine-tunes the full pipeline on a smaller target dataset. This approach delivers state-of-the-art per-object mask accuracy with a fraction of the labeled data and compute that training from scratch would require.