OpenCodePapers

open-vocabulary-semantic-segmentation-on-ade20k-150

Semantic SegmentationOpen Vocabulary Semantic Segmentation
Dataset Link
Results over time
Click legend items to toggle metrics. Hover points for model names.
Leaderboard
PaperCodemIoUModelNameReleaseDate
dinov3.seg: Open-Vocabulary Semantic Segmentation with DINOv342.19dinov3.seg2026-03-19
PCA-Seg: Revisiting Cost Aggregation for Open-Vocabulary Semantic and Part Segmentation✓ Link41.5PCA-Seg (ViT-L/14)2026-03-18
Reconstructing CLIP for Open-Vocabulary Dense Perception41.5DenseRC (ViT-L/14)2026-06-03
EOVSAM: Efficient Open-Vocabulary Segmentation with SAM 3 in One Pass39.0EOVSAM2026-08-03
DiSa: Saliency-Aware Foreground-Background Disentangled Framework for Open-Vocabulary Semantic Segmentation38.9DiSa (CLIP ViT-L/14)2026-01-27
SigLIP 2: Multilingual Vision-Language Encoders with Improved Semantic Understanding, Localization, and Dense Features✓ Link38.8SigLIP 2 (L/16) + Cat-Seg2025-02-20
HyperCLIP++: Fine-tuning CLIP forOpen-vocabulary Semantic Segmentation in Hyperbolic Space38.7HyperCLIP++ (CLIP ViT-L/14)2026-09-21
InfoCLIP: Bridging Vision-Language Pretraining and Open-Vocabulary Semantic Segmentation via Information-Theoretic Alignment Transfer38.5InfoCLIP (CLIP ViT-L/14)2025-11-20
UMG-CLIP: A Unified Multi-Granularity Vision Generalist for Open-World Understanding✓ Link38.2UMG-CLIP-E/142024-01-12
Mask-Adapter: The Devil is in the Masks for Open-Vocabulary Segmentation✓ Link38.2Mask-Adapter2024-12-05
MaskCLIP++: A Mask-Based CLIP Fine-tuning Framework for Open-Vocabulary Image Segmentation✓ Link38.2MaskCLIP++2024-12-16
Understanding Fine-tuning CLIP for Open-vocabulary Semantic Segmentation in Hyperbolic Space✓ Link38.2HyperCLIP (ViT-L/14)2025-06-11
Novel Category Discovery with X-Agent Attention for Open-Vocabulary Semantic Segmentation38.2X-Agent (CLIP ViT-L/14)2025-09-01
DCP-CLIP:A Coarse-to-Fine Framework for Open-Vocabulary Semantic Segmentation with Dual Interaction38.1DCP-CLIP (CLIP ViT-L/14)2026-03-14
PCA-Seg: Revisiting Cost Aggregation for Open-Vocabulary Semantic and Part Segmentation✓ Link38.1PCA-Seg (ViT-B/16)2026-03-18
CAT-Seg: Cost Aggregation for Open-Vocabulary Semantic Segmentation✓ Link37.9CAT-Seg2023-03-21
SILC: Improving Vision Language Pretraining with Self-Distillation37.7SILC2023-10-20
UMG-CLIP: A Unified Multi-Granularity Vision Generalist for Open-World Understanding✓ Link36.1UMG-CLIP-L/142024-01-12
Collaborative Vision-Text Representation Optimizing for Open-Vocabulary Segmentation✓ Link36.1MAFT+2024-08-01
OpenDAS: Open-Vocabulary Domain Adaptation for 2D and 3D Segmentation35.8OVSeg + OpenDAS2024-05-30
SED: A Simple Encoder-Decoder for Open-Vocabulary Semantic Segmentation✓ Link35.2SED2023-11-27
CLIPSelf: Vision Transformer Distills Itself for Open-Vocabulary Dense Prediction✓ Link34.5CLIPSelf2023-10-02
Convolutions Die Hard: Open-Vocabulary Segmentation with Single Frozen Convolutional CLIP✓ Link34.1FC-CLIP2023-08-04
DiSa: Saliency-Aware Foreground-Background Disentangled Framework for Open-Vocabulary Semantic Segmentation33.7DiSa (CLIP ViT-B/16)2026-01-27
Open-Vocabulary Segmentation with Semantic-Assisted Calibration✓ Link33.5SCAN2023-12-07
HyperCLIP++: Fine-tuning CLIP forOpen-vocabulary Semantic Segmentation in Hyperbolic Space32.9HyperCLIP++ (CLIP ViT-B/16)2026-09-21
Open-Vocabulary Semantic Segmentation with Image Embedding Balancing✓ Link32.8EBSeg-L2024-06-14
DCP-CLIP:A Coarse-to-Fine Framework for Open-Vocabulary Semantic Segmentation with Dual Interaction32.4DCP-CLIP (CLIP ViT-B/16)2026-03-14
Novel Category Discovery with X-Agent Attention for Open-Vocabulary Semantic Segmentation32.1X-Agent (CLIP ViT-B/16)2025-09-01
InfoCLIP: Bridging Vision-Language Pretraining and Open-Vocabulary Semantic Segmentation via Information-Theoretic Alignment Transfer32.1InfoCLIP (CLIP ViT-B/16)2025-11-20
Learning Mask-aware CLIP Representations for Zero-Shot Segmentation✓ Link32.0MAFT-ViTL2023-09-30
Open Vocabulary Semantic Segmentation with Patch Aligned Contrastive Learning✓ Link31.4PACL2022-12-09
Open-Vocabulary Panoptic Segmentation with Text-to-Image Diffusion Models✓ Link29.9ODISE2023-03-08
Open-Vocabulary Semantic Segmentation with Mask-adapted CLIP✓ Link29.6OVSeg Swin-B2022-10-09
Open-Vocabulary Universal Image Segmentation with MaskCLIP✓ Link23.7MaskCLIP2022-08-18
20.7POMP
A Simple Baseline for Open-Vocabulary Semantic Segmentation with Pre-trained Vision-language Model✓ Link20.5SimSeg2021-12-29
TTD: Text-Tag Self-Distillation Enhancing Image-Text Alignment in CLIP to Alleviate Single Tag Bias✓ Link17.0TTD (TCL)2024-03-30
In Defense of Lazy Visual Grounding for Open-Vocabulary Semantic Segmentation✓ Link15.8LaVG2024-08-09
TTD: Text-Tag Self-Distillation Enhancing Image-Text Alignment in CLIP to Alleviate Single Tag Bias✓ Link12.7TTD (MaskCLIP)2024-03-30