SAM segment anything
Meta:segment anything
介绍地址：https://ai.facebook.com/research/publications/segment-anything/
演示地址：https://segment-anything.com/demo#
论文地址：https://arxiv.org/abs/2304.02643
GitHub地址：https://github.com/facebookresearch/segment-anything
Abstract
We introduce the Segment Anything (SA) project: a new task, model, and dataset for image segmentation. Using our efficient model in a data collection loop, we built the largest segmentation dataset to date (by far), with over 1 billion masks on 11M licensed and privacy respecting images. The model is designed and trained to be promptable, so it can transfer zero-shot to new image distributions and tasks. We evaluate its capabilities on numerous tasks and find that its zero-shot performance is impressive – often competitive with or even superior to prior fully supervised results. We are releasing the Segment Anything Model (SAM) and corresponding dataset (SA-1B) of 1B masks and 11M images at \href{https://segment-anything.com}{https://segment-anything.com} to foster research into foundation models for computer vision.
SEEM: Segment Everything Everywhere All At Once
[外链图片转存失败,源站可能有防盗链机制,建议将图片保存下来直接上传(img-E9i81rhc-1686562293452)(https://files.mdnice.com/user/44170/30a091e7-8ab4-4dc7-9b00-5d275d504d95.png)]
地址：https://huggingface.co/spaces/xdecoder/SEEM
Instructions:
Try our default examples first (Sketch is not automatically drawed on input and example image);
For video demo, it takes about 30-60s to process, please refresh if you meet an error on uploading;
Upload an image/video (If you want to use referred region of another image please check “Example” and upload another image in referring image panel);
Select at least one type of prompt of your choice (If you want to use referred region of another image please check “Example”);
Remember to provide the actual prompt for each promt type you select, otherwise you will meet an error (e.g., rember to draw on the referring image);
Our model by defualt support the vocabulary of COCO 133 categories, others will be classified to ‘others’ or misclassifed.
SEEM：微软基于 CV 大模型新作，分割“瞬息全宇宙”
论文题目:
Segment Everything Everywhere All at Once
论文链接:
https://arxiv.org/abs/2304.06718
项目地址:
https://github.com/ux-decoder/segment-everything-everywhere-all-at-once
Demo地址:
https://36771ee9c49a4631.gradio.app/
图像分割：SAM 与 SEEM
在分割问题领域，Meta 几天前提出的 SAM 提供了一个通用且全自动的图像分割方法，它的创新之处在于可以同时执行交互式分割和自动分割，并且可以通过灵活的 prompt 界面来适应新任务和新领域。它解决了传统方法需要很多手动注释和对于特定对象的限制的问题，具有很高的适用性和可扩展性。
本文由mdnice多平台发布