The field of computer vision is witnessing significant advancements in video object segmentation and medical image analysis. Researchers are exploring innovative approaches to improve the accuracy and efficiency of models in these areas. One notable direction is the development of foundation models that can be fine-tuned for specific tasks, such as segmenting objects in surgical videos or analyzing medical images. Another area of focus is the use of self-supervised learning techniques to learn visual representations from diverse sources, eliminating the need for manual data annotation. These advancements have the potential to revolutionize various applications, including medical diagnostics and robotic surgery. Noteworthy papers in this area include: TSMS-SAM2, which introduces a novel framework for promptable video object segmentation and tracking in surgical scenarios, and RedDino, a self-supervised foundation model designed for red blood cell image analysis. S2-UniSeg is also notable for its fast universal agglomerative pooling approach for scalable segment anything without supervision. DINOv3 is a major milestone in self-supervised learning, leveraging simple yet effective strategies to learn visual representations from diverse sources.