mlsanityData to 7 Oct 2026

Segment Anything in High Quality

arXiv 2306.01567 ↗First posted 2 Jun 2023cs.CVcited by 373

Not on the map. Why?

Abstract (arXiv)

The recent Segment Anything Model (SAM) represents a big leap in scaling up segmentation models, allowing for powerful zero-shot capabilities and flexible prompting. Despite being trained with 1.1 billion masks, SAM's mask prediction quality falls short in many cases, particularly when dealing with objects that have intricate structures. We propose HQ-SAM, equipping SAM with the ability to accurately segment any object, while maintaining SAM's original promptable design, efficiency, and zero-shot generalizability. Our careful design reuses and preserves the pre-trained model weights of SAM, while only introducing minimal additional parameters and computation. We design a learnable High-Quality Output Token, which is injected into SAM's mask decoder and is responsible for predicting the high-quality mask. Instead of only applying it on mask-decoder features, we first fuse them with early and final ViT features for improved mask details. To train our introduced learnable parameters, we compose a dataset of 44K fine-grained masks from several sources. HQ-SAM is only trained on the introduced detaset of 44k masks, which takes only 4 hours on 8 GPUs. We show the efficacy of HQ-SAM in a suite of 10 diverse segmentation datasets across different downstream tasks, where 8 out of them are evaluated in a zero-shot transfer protocol. Our code and pretrained models are at https://github.com/SysCV/SAM-HQ.

Key facts. Cited by 373 papers in the map (data through 7 Oct 2026). Builds on 50 of its 59 references that are in the map. First posted on arXiv 2 Jun 2023.

Builds on

50 of 59 references.

References · 50

  1. 2021Learning Transferable Visual Models From Natural Language Supervisioncited by 33,456
  2. 2020Language Models are Few-Shot Learnerscited by 32,199
  3. 2020An Image is Worth 16x16 Words: Transformers for Image Recognition at Scalecited by 29,915
  4. 2014Microsoft COCO: Common Objects in Contextcited by 17,498
  5. 2021LoRA: Low-Rank Adaptation of Large Language Modelscited by 14,084
  6. 2023Segment Anythingcited by 8,158
  7. 2020End-to-End Object Detection with Transformerscited by 7,384
  8. 2019Parameter-Efficient Transfer Learning for NLPcited by 4,195
  9. 2021On the Opportunities and Risks of Foundation Modelscited by 3,961
  10. 2021Prefix-Tuning: Optimizing Continuous Prompts for Generationcited by 3,878
  11. 2023Grounding DINO: Marrying DINO with Grounded Pre-Training for Open-Set Object Detectioncited by 3,274
  12. 2016Pyramid Scene Parsing Networkcited by 2,472
38 more
  1. 2021Learning to Prompt for Vision-Language Modelscited by 2,269
  2. 2022Visual Prompt Tuningcited by 1,674
  3. 2022DINO: DETR with Improved DeNoising Anchor Boxes for End-to-End Object Detectioncited by 1,306
  4. 2019LVIS: A Dataset for Large Vocabulary Instance Segmentationcited by 1,215
  5. 2017The 2017 DAVIS Challenge on Video Object Segmentationcited by 942
  6. 2022Exploring Plain Vision Transformer Backbones for Object Detectioncited by 723
  7. 2020Simple Copy-Paste is a Strong Data Augmentation Method for Instance Segmentationcited by 595
  8. 2023Segment Everything Everywhere All at Oncecited by 546
  9. 2022DN-DETR: Accelerate DETR Training by Introducing Query DeNoisingcited by 486
  10. 2016RefineNet: Multi-Path Refinement Networks for High-Resolution Semantic Segmentationcited by 463
  11. 2019PointRend: Image Segmentation as Renderingcited by 439
  12. 2022XMem: Long-Term Video Object Segmentation with an Atkinson-Shiffrin Memory Modelcited by 436
  13. 2019Video Instance Segmentationcited by 391
  14. 2022Generalized Decoding for Pixel, Image, and Languagecited by 292
  15. 2017ICNet for Real-Time Semantic Segmentation on High-Resolution Imagescited by 261
  16. 2023SegGPT: Segmenting Everything In Contextcited by 193
  17. 2021Boundary IoU: Improving Object-Centric Image Segmentation Evaluationcited by 186
  18. 2021CPT: Colorful Prompt Tuning for Pre-trained Vision-Language Modelscited by 182
  19. 2022ZegCLIP: Towards Adapting CLIP for Zero-shot Semantic Segmentationcited by 178
  20. 2021Mask2Former for Video Instance Segmentationcited by 169
  21. 2019Gated-SCNN: Gated Shape CNNs for Semantic Segmentationcited by 165
  22. 2019FSS-1000: A 1000-Class Dataset for Few-Shot Segmentationcited by 160
  23. 2023AIM: Adapting Image Models for Efficient Video Action Recognitioncited by 145
  24. 2021Unidentified Video Objects: A Benchmark for Dense, Open-World Segmentationcited by 133
  25. 2020CascadePSP: Toward Class-Agnostic and Very High-Resolution Segmentation via Global and Local Refinementcited by 113
  26. 2022Highly Accurate Dichotomous Image Segmentationcited by 110
  27. 2020SegFix: Model-Agnostic Boundary Refinement for Segmentationcited by 97
  28. 2019Towards High-Resolution Salient Object Detectioncited by 92
  29. 2022What do Vision Transformers Learn? A Visual Explorationcited by 75
  30. 2021Mask Transfiner for High-Quality Instance Segmentationcited by 64
  31. 2017On Pre-Trained Image Features and Synthetic Images for Deep Learningcited by 51
  32. 2021Look Closer to Segment Better: Boundary Patch Refinement for Instance Segmentationcited by 43
  33. 2021High Quality Segmentation for Ultra High-resolution Imagescited by 31
  34. 2022Video Mask Transfiner for High-Quality Video Instance Segmentationcited by 28
  35. 2020NDD20: A large-scale few-shot dolphin dataset for coarse and fine-grained categorisationcited by 24
  36. 2018Semantic Segmentation Refinement by Monte Carlo Region Growing of High Confidence Detectionscited by 14
  37. 2023PatchDCT: Patch Refinement for High Quality Instance Segmentationcited by 8
  38. 2023Intra-Batch Supervision for Panoptic Segmentation on High-Resolution Imagescited by 3

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2023: 45; 2024: 139; 2025: 116; 2026: 73.

373 papers.

Sort:

By area:

  1. 2024SAM 2: Segment Anything in Images and Videoscited by 2,784
  2. 2024Depth Anything: Unleashing the Power of Large-Scale Unlabeled Datacited by 1,346
  3. 2024Grounded SAM: Assembling Open-World Models for Diverse Visual Taskscited by 914
  4. 2023GLaMM: Pixel Grounding Large Multimodal Modelcited by 447
  5. 2023Gaussian Grouping: Segment and Edit Anything in 3D Scenescited by 387
  6. 2024SpatialRGPT: Grounded Spatial Reasoning in Vision Language Modelscited by 297
  7. 2023Tracking Anything with Decoupled Video Segmentationcited by 197
  8. 2023RSPrompter: Learning to Prompt for Remote Sensing Instance Segmentation based on Visual Foundation Modelcited by 180
  9. 2023Osprey: Pixel Understanding with Visual Instruction Tuningcited by 158
  10. 2023RepViT: Revisiting Mobile CNN From ViT Perspectivecited by 123
  11. 2023Parameter-Efficient Orthogonal Finetuning via Butterfly Factorizationcited by 112
  12. 2023SAM-CLIP: Merging Vision Foundation Models towards Semantic and Spatial Understandingcited by 111
38 more
  1. 2023SAI3D: Segment Any Instance in 3D Scenescited by 90
  2. 2025MedSAM2: Segment Anything in 3D Medical Images and Videoscited by 88
  3. 2023Segment Anything Meets Point Trackingcited by 81
  4. 2024VRP-SAM: SAM with Visual Reference Promptcited by 75
  5. 2024VideoRefer Suite: Advancing Spatial-Temporal Object Understanding with Video LLMcited by 71
  6. 2023CLIP as RNN: Segment Countless Visual Concepts without Training Endeavorcited by 69
  7. 2024DreamScene4D: Dynamic Multi-Object Scene Generation from Monocular Videoscited by 63
  8. 2024EVF-SAM: Early Vision-Language Fusion for Text-Prompted Segment Anything Modelcited by 63
  9. 2024ParaHome: Parameterizing Everyday Home Activities Towards 3D Generative Modeling of Human-Object Interactionscited by 54
  10. 2024LLM-Seg: Bridging Image Segmentation and Large Language Model Reasoningcited by 52
  11. 2024Matching Anything by Segmenting Anythingcited by 49
  12. 2024Watermark Anything with Localized Messagescited by 49
  13. 2025Perceive Anything: Recognize, Explain, Caption, and Segment Anything in Images and Videoscited by 49
  14. 2024RoboEXP: Action-Conditioned Scene Graph via Interactive Exploration for Robotic Manipulationcited by 48
  15. 2023Synthesize, Diagnose, and Optimize: Towards Fine-Grained Vision-Language Understandingcited by 46
  16. 2024DreamStory: Open-Domain Story Visualization by LLM-Guided Multi-Subject Consistent Diffusioncited by 46
  17. 2024IRSAM: Advancing Segment Anything Model for Infrared Small Target Detectioncited by 45
  18. 2023SAMPro3D: Locating SAM Prompts in 3D for Zero-Shot Instance Segmentationcited by 44
  19. 2024Segment Anything in Medical Images and Videos: Benchmark and Deploymentcited by 42
  20. 2023MobileSAMv2: Faster Segment Anything to Everythingcited by 41
  21. 2025PosterMaker: Towards High-Quality Product Poster Generation with Accurate Text Renderingcited by 41
  22. 2023TokenCompose: Text-to-Image Diffusion with Token-level Supervisioncited by 40
  23. 2024MULAN: A Multi Layer Annotated Dataset for Controllable Text-to-Image Generationcited by 38
  24. 2025OmniPaint: Mastering Object-Oriented Editing via Disentangled Insertion-Removal Inpaintingcited by 38
  25. 2023Template Free Reconstruction of Human-object Interaction with Procedural Interaction Generationcited by 37
  26. 2024RobustSAM: Segment Anything Robustly on Degraded Imagescited by 36
  27. 2023SAM-guided Graph Cut for 3D Instance Segmentationcited by 35
  28. 2024Image Segmentation in Foundation Model Era: A Surveycited by 35
  29. 2024CorrCLIP: Reconstructing Patch Correlations in CLIP for Open-Vocabulary Semantic Segmentationcited by 33
  30. 2024PTQ4SAM: Post-Training Quantization for Segment Anythingcited by 32
  31. 2024DarkSAM: Fooling Segment Anything Model to Segment Nothingcited by 32
  32. 2025Vid2Avatar-Pro: Authentic Avatar from Videos in the Wild via Universal Priorcited by 31
  33. 2024Hi-SAM: Marrying Segment Anything Model for Hierarchical Text Segmentationcited by 30
  34. 2024Interactive Medical Image Segmentation: A Benchmark Dataset and Baselinecited by 30
  35. 2025Vid2Sim: Realistic and Interactive Simulation from Video for Urban Navigationcited by 30
  36. 2024SAGD: Boundary-Enhanced Segment Anything in 3D Gaussian via Gaussian Decompositioncited by 29
  37. 2025SegAgent: Exploring Pixel Understanding Capabilities in MLLMs by Imitating Human Annotator Trajectoriescited by 29
  38. 2024CAT-SAM: Conditional Tuning for Few-Shot Adaptation of Segment Anything Modelcited by 27

50 most cited of 373 All 373 (JSON)