FrontVeg V2: A Training-Free Software Framework for Foreground-Aware Zero-Shot Plant Trait Segmentation in High-Resolution Images of Trellised Crops

Abdoul Djalil Ousseini Hamza, Herearii Metuarea, Corentin Lothod{é}, Morgane Roth, Jacem Ben Hamden +4 more
10/5/2026
cs.CV

No official code found. Best implementation: djaliloh/FrontVegV2 (6 stars, MIT, last commit 2026-09-28).

Abstract

FrontVeg V2 is an open-source, training-free software framework for foregroundaware zero-shot segmentation of plant traits in high-resolution images of trellised crops. The pipeline combines monocular depth estimation, automatic foreground extraction using Valley-Aware Depth Thresholding, tiled zero-shot segmentation, Graph-Based Mask Assembly, and geometry-aware fusion. This design enables plant organs and disease symptoms to be segmented while reducing detections arising from neighboring vegetation rows. The current implementation integrates Depth Anything V2 (DAV2) and SAM3 and can be used through both command-line batch processing and a Napari graphical interface. FrontVeg V2 provides a reusable framework for multi-crop, multi-trait digital phenotyping without task-specific model retraining.

View on arXivView PDF

Code Implementations(1)

djaliloh/FrontVegV2Community71%
MIT

[FrontVeg V2] Official implementation of our paper: "FrontVeg V2: A Training-Free Software Framework for Foreground-Aware Zero-Shot Plant Trait Segmentation in High-Resolution Images of Trellised Crops"

61Nov 7, 20243 weeks agoMIT
foundation-modelsmonocular-depth-estimationplant-phenotypingtrellised-cropsvalley-aware-depth-thresholding+1 more

Similar Papers

Unlocking Zero-Shot Plant Segmentation with Pl@ntNet Intelligence

83% match2025

Zero-shot Hierarchical Plant Segmentation via Foundation Segmentation Models and Text-to-image Attention

83% match1 repo72025

Multi-Scale ViT Inference with Habitat-Fit Priors and kNN Retrieval for Multi-Species Plant Identification

81% match1 repo22026

IPENS:Interactive Unsupervised Framework for Rapid Plant Phenotyping Extraction via NeRF-SAM2 Fusion

81% match2025

Enhancing Ecological Monitoring with Multi-Objective Optimization: A Novel Dataset and Methodology for Segmentation Algorithms

81% match2024

The Power of Transfer Learning in Agricultural Applications: AgriNet

80% match2022

Cite this paper

@article{hamza2026frontveg,
  title  = {FrontVeg V2: A Training-Free Software Framework for Foreground-Aware Zero-Shot Plant Trait Segmentation in High-Resolution Images of Trellised Crops},
  author = {Abdoul Djalil Ousseini Hamza and Herearii Metuarea and Corentin Lothod{é} and Morgane Roth and Jacem Ben Hamden and Eric Duch{ê}ne and Lionel Ley and David Alletru and David Rousseau},
  year   = {2026},
  eprint = {2610.06575},
  archivePrefix = {arXiv},
  url    = {http://arxiv.org/abs/2610.06575v1}
}

Add a badge to your README

Shows whether official code is linked for this paper.

Code of Paper badge
[![Code of Paper](https://codeofpaper.com/badge/2610.06575.svg)](https://codeofpaper.com/paper/2610.06575)

Discussion