Skip to content
Preprint

ProGuT: Label-Efficient Panoptic Segmentation for Forest Scenes

Sep 2026 · 0 citations · 39 references
Computer Science

TL;DR

ProGuT (Prototype Guided Training), which produces panoptic pseudo-labels without per-image training masks, needing only unlabeled images and one-time cluster-to-class mapping, outperforms existing unsupervised methods for class-agnostic trunk instance benchmark.

Abstract

Panoptic segmentation in forest environments is bottlenecked not by semantic quality but by instance separation; existing unsupervised panoptic approaches produce usable stuff maps but near-zero thing quality. Depth or flow-based instance discovery methods needs sensors that are not always available. We present ProGuT (Prototype Guided Training), which produces panoptic pseudo-labels without per-image training masks, needing only unlabeled images and one-time cluster-to-class mapping. ProGuT clusters CLIP patch features, then recovers trunk instances through multiscale geometric prior that falsifies non-trunk structures via structure-tensor. This is cheap compared to depth, flow or class-supervision methods to create pseudo labels. These are then used for downstream tasks which we evaluate against other unsupervised baselines. ProGuT achieves a Panoptic Quality (PQ) of 65.2 on Our-forest dataset (2.6x improvement over the initial pseudo-label quality) and reaches 65.9 mIoU on Freiburg Forest, outperforming unsupervised baselines like PiCIE (45.3 IoU) and STEGO(57.6IoU). Additionally, ProGuT outperforms existing unsupervised methods for class-agnostic trunk instance benchmark.

View source

Similar papers

Open access Sep 2026

MC-PanDA++: Simpler, Stronger, and More Robust Domain-Adaptive Panoptic Segmentation

Unsupervised domain adaptation (UDA) reduces the annotation burden in panoptic segmentation by leveraging a cost-effectively labeled source domain (e.g., synthetic) and an unlabeled target domain to bridge the distribution gap. Existing panoptic UDA methods rely on teacher-student consistency learning built upon subopt...

Ivan Martinović, Josip Saric, Yuki M. Asano et al. · 0 citations
Preprint Oct 2026

ZeroDSS: Semantic Segmentation from Class Names Alone

Semantic segmentation depends heavily on real images with pixel-level annotations, making dense supervision a major obstacle to deployment in new domains. We study task-specific data-free semantic segmentation, where training uses generic pretrained generative priors but has access to no target-domain images, masks, or...

Ming Tian, Marzi Heidari, Yu-Hong Guo · 0 citations
Open access 2026

Dual-Level Prototype Alignment via Cross-Attention for Few-Shot Remote Sensing Semantic Segmentation

DLPANet is proposed, a novel dual-level prototype alignment network centered on Prototype-Guided Spatial Attention, enabling simultaneous modeling of scene context and fine-grained details and demonstrates that the decoupled dual cross-attention mechanism provides superior prototype-query alignment compared to prior gl...

Mustafa Alawadi, M. Fateh · 0 citations
Preprint Oct 2026

TAG: Calibrating SAM3 for Universal Image Segmentation without Dense Supervision

The recent Segment Anything Model 3 (SAM3) extends powerful promptable segmentation from spatial cues to concepts, raising the prospect of a single foundation model for universal image segmentation. However, this promise often fails to translate into reliable performance on downstream segmentation tasks. In this work,...

Hao Zhu, He Sun, Shuo Jin et al. · 0 citations
#machine learning Preprint Sep 2026

Panoptic Scene Program Diffusion Transformer

Under matched training and inference settings, PSP-DiT improves over a strong flat-text baseline across GenEval 2, SANEval-Simple, PSG-Score, and DetailMaster, with the largest gains on counting, attribute binding, role-sensitive relations, and long structured prompts.

C. Maduabuchi · 0 citations
Open access Oct 2026

DINOSim: Training-Free Semantic Segmentation on Microscopy Images.

We present DINOSim, a training-free method for segmenting objects in microscopy images without the need for large annotated datasets or additional model training. DINOSim builds on the pretrained DINOv2 image encoder, which captures semantic information from images. By comparing the encoder's features of image patches...

Aitor González-Marfil, Estibaliz Gómez-de-Mariscal, I. Arganda-Carreras · 0 citations

We use cookies to run the site and, with your consent, for analytics and to show ads. See our Cookie Policy.