Papers › A Simple Baseline for Open-Vocabulary Semantic Segmentation with Pre-trained...

A Simple Baseline for Open-Vocabulary Semantic Segmentation with Pre-trained Vision-language Model

29 Dec 2021arXiv:2112.14757archive 2025-07-28

Mengde Xu, Zheng Zhang, Fangyun Wei, Yutong Lin, Yue Cao, Han Hu, Xiang Bai

Recently, open-vocabulary image classification by vision language pre-training has demonstrated incredible achievements, that the model can classify arbitrary categories without seeing additional annotated images of that category. However, it is still unclear how to make the open-vocabulary recognition work well on broader vision problems. This paper targets open-vocabulary semantic segmentation by building it on an off-the-shelf pre-trained vision-language model, i.e., CLIP. However, semantic segmentation and the CLIP model perform on different visual granularity, that semantic segmentation processes on pixels while CLIP performs on images. To remedy the discrepancy in processing granularity, we refuse the use of the prevalent one-stage FCN based framework, and advocate a two-stage semantic segmentation framework, with the first stage extracting generalizable mask proposals and the second stage leveraging an image based CLIP model to perform open-vocabulary classification on the masked image crops which are generated in the first stage. Our experimental results show that this two-stage framework can achieve superior performance than FCN when trained only on COCO Stuff dataset and evaluated on other datasets without fine-tuning. Moreover, this simple framework also surpasses previous state-of-the-arts of zero-shot semantic segmentation by a large margin: +29.5 hIoU on the Pascal VOC 2012 dataset, and +8.9 hIoU on the COCO Stuff dataset. With its simplicity and strong performance, we hope this framework to serve as a baseline to facilitate future research. The code are made publicly available at~\url{https://github.com/MendelXu/zsseg.baseline}.

PaperPDFCode

In Syntology Open this paper in Syntology's Atlas, the map of the papers in Syntology's graph and their citations.

Code

mendelxu/zsseg.baseline officialmentioned in papermentioned on GitHubpytorchMIT report
openrobotlab/ov_parts mentioned on GitHubjax report

Repository list and official/mentioned flags are the archive's, frozen 2025-07-28. Reachability, where shown, is from one Syntology probe window (2026-09-16 to 2026-09-18); repositories not probed show nothing. GitHub stars are not tracked.

Code Syntology ran Syntology

Not run by Syntology. Nothing on this page verifies that the listed code works.

Tasks

Image ClassificationLanguage ModelingLanguage ModellingObject DetectionOpen Vocabulary Image ClassificationOpen Vocabulary Semantic SegmentationOpen-Vocabulary Semantic SegmentationSegmentationSemantic SegmentationZero-Shot Image ClassificationZero-Shot LearningZero-Shot Semantic Segmentationimage-classificationobject-detection

Results from the paper archive 2025-07-28

TaskDatasetModelMetricValueRank at snapshotLeaderboardReport
Open Vocabulary Semantic Segmentation ADE20K-150 SimSeg mIoU 20.5 #20 of 23 Archive leaderboard report
Open Vocabulary Semantic Segmentation ADE20K-847 SimSeg mIoU 7 #19 of 19 Archive leaderboard report
Open Vocabulary Semantic Segmentation COCO-Stuff-171 ZSSeg HIoU 37.8 #2 of 7 Archive leaderboard report
Open Vocabulary Semantic Segmentation Cityscapes SimSeg mIoU 34.5 #2 of 5 Archive leaderboard report
Open Vocabulary Semantic Segmentation PASCAL Context-59 SimSeg mIoU 47.7 #17 of 24 Archive leaderboard report
Open Vocabulary Semantic Segmentation PascalVOC-20 ZSSeg hIoU 77.5 #19 of 20 Archive leaderboard report
Zero-Shot Semantic Segmentation COCO-Stuff zsseg Inductive Setting hIoU 36.3 #8 of 15 Archive leaderboard report
Zero-Shot Semantic Segmentation COCO-Stuff zsseg Transductive Setting hIoU 41.5 #8 of 15 Archive leaderboard report
Zero-Shot Semantic Segmentation PASCAL VOC zsseg Inductive Setting hIoU 77.5 #7 of 13 Archive leaderboard report
Zero-Shot Semantic Segmentation PASCAL VOC zsseg Transductive Setting hIoU 79.3 #7 of 13 Archive leaderboard report

Ranks are positions in the archive's leaderboards as they stood at the 2025-07-28 snapshot. Results published since then are not among these rows, so a rank here is not a current standing.

Methods

CLIPConvolutionFCNMax Pooling

Report a problem or propose a change · a person checks every report against the paper or source before anything changes; decisions are listed on /corrections