使用 Mask2Former 與 OneFormer 進行通用影像分割
Original: Universal Image Segmentation with Mask2Former and OneFormer
Image segmentation is a core task in computer vision, traditionally divided into three main types: semantic segmentation (classifying every…
Hugging Face 宣布在 transformers 庫中支援 Mask2Former 與 OneFormer 兩大通用影像分割模型。這兩款模型打破了以往語意、實例和全景分割需要不同架構的限制，實現「單一架構通吃所有分割任務」。開發者現在可以透過簡單的 API 輕鬆載入預訓練模型，並應用於各類電腦視覺場景。
Image segmentation is a core task in computer vision, traditionally divided into three main types: semantic segmentation (classifying every pixel), instance segmentation (detecting and distinguishing individual objects), and panoptic segmentation (combining the two). In the past, each of these three tasks typically required its own dedicated and complex model architecture.
Free shows the 3-line summary; Pro unlocks the full deep summary (~300 words) so you never have to click through.
See Pro plans →Want the original English / full article?
Read on Hugging Face Blog →Summaries are AI-generated; the original article is authoritative.