AI Paper Atlas

Research map / Detection and segmentation / Open-set and few-shot detection

ProSAM: Enhancing the Robustness of SAM-based Visual Reference Segmentation with Probabilistic Prompts

Xiaoqi Wang, Clint Sebastian, Wenbin He, Liu Ren

ICCV 2025 · Detection and segmentation › Open-set and few-shot detection · approach: Prompted SAM segmentation · 6 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
replace VRP-SAM prompt encoder with variational t-distribution prompt encoderVRP-SAM · COCO-20i (mask)mIoU: 53.28 -> 54.63Improved
replace VRP-SAM prompt encoder with variational prompt encoderVRP-SAM · PASCAL-5i (mask)mIoU: 71.14 -> 72.04Improved
use Gaussian instead of heavy-tailed t prompt distribution (ablation)ProSAM · PASCAL-5i (mask)mIoU: 72.04 -> 71.41Hurt
increase Monte Carlo samples K from 10 to 15 (ablation)ProSAM · PASCAL-5i (mask)mIoU: 72.04 -> 71.42Hurt
reduce t-distribution degrees of freedom from 5 to 3 (ablation)ProSAM · PASCAL-5i (mask)mIoU: 72.04 -> 71.55Hurt
append two extra linear layers to VRP-SAM to match parameters (ablation)VRP-SAM · PASCAL-5i F-0mIoU: 74.01 -> 74.04No clear effect
swap ResNet-50 for DINOv2 ViT-B/14 image encoder (ablation)ProSAM · COCO-20i (mask)mIoU: 54.63 -> 59.16Improved
swap ResNet-50 for VGG-16 image encoder (ablation)ProSAM · PASCAL-5i (mask)mIoU: 72.04 -> 69.27Hurt
use sampled prompts with K masks merged instead of mean prompt at inference (ablation)ProSAM · PASCAL-5i F-0mIoU: 75.26 -> 75.16No clear effect
add Gaussian noise (std 1.2) to prompt embeddings at inferenceVRP-SAM vs ProSAM · PASCAL-5i F-0mIoU degradation: VRP-SAM about 40% drop vs ProSAM about 20% dropImproved
use variational prompt encoder under COCO to PASCAL domain shiftVRP-SAM · COCO-20i to PASCAL-5imean mIoU: 76.44 -> 77.65Improved

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy