Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

LeafInst - Unified Instance Segmentation Network for Fine-Grained Forestry Leaf Phenotype Analysis: A New UAV based Benchmark

Taige Luo, Junru Xie, Chenyang Fan, Bingrong Liu, Ruisheng Wang, Yang Shao, Sheng Xu, Lin Cao

TL;DR

This work proposes LeafInst, a novel segmentation framework tailored for irregular and multi-scale leaf structures, which integrates an Asymptotic Feature Pyramid Network for multi-scale perception, a Dynamic Asymmetric Spatial Perception module for irregular shape modeling, and a dual-residual Dynamic Anomalous Regression Head with Top-down Concatenation decoder Feature Fusion to improve detection and segmentation performance.

Abstract

Intelligent forest tree breeding has advanced plant phenotyping, yet existing research largely focuses on large-leaf agricultural crops, with limited attention to fine-grained leaf analysis of sapling trees in open-field environments. Natural scenes introduce challenges including scale variation, illumination changes, and irregular leaf morphology. To address these issues, we collected UAV RGB imagery of field-grown saplings and constructed the Poplar-leaf dataset, containing 1,202 branches and 19,876 pixel-level annotated leaf instances. To our knowledge, this is the first instance segmentation dataset specifically designed for forestry leaves in open-field conditions. We propose LeafInst, a novel segmentation framework tailored for irregular and multi-scale leaf structures. The model integrates an Asymptotic Feature Pyramid Network (AFPN) for multi-scale perception, a Dynamic Asymmetric Spatial Perception (DASP) module for irregular shape modeling, and a dual-residual Dynamic Anomalous Regression Head (DARH) with Top-down Concatenation decoder Feature Fusion (TCFU) to improve detection and segmentation performance. On Poplar-leaf, LeafInst achieves 68.4 mAP, outperforming YOLOv11 by 7.1 percent and MaskDINO by 6.5 percent. On the public PhenoBench benchmark, it reaches 52.7 box mAP, exceeding MaskDINO by 3.4 percent. Additional experiments demonstrate strong generalization and practical utility for large-scale leaf phenotyping.

LeafInst - Unified Instance Segmentation Network for Fine-Grained Forestry Leaf Phenotype Analysis: A New UAV based Benchmark

TL;DR

This work proposes LeafInst, a novel segmentation framework tailored for irregular and multi-scale leaf structures, which integrates an Asymptotic Feature Pyramid Network for multi-scale perception, a Dynamic Asymmetric Spatial Perception module for irregular shape modeling, and a dual-residual Dynamic Anomalous Regression Head with Top-down Concatenation decoder Feature Fusion to improve detection and segmentation performance.

Abstract

Intelligent forest tree breeding has advanced plant phenotyping, yet existing research largely focuses on large-leaf agricultural crops, with limited attention to fine-grained leaf analysis of sapling trees in open-field environments. Natural scenes introduce challenges including scale variation, illumination changes, and irregular leaf morphology. To address these issues, we collected UAV RGB imagery of field-grown saplings and constructed the Poplar-leaf dataset, containing 1,202 branches and 19,876 pixel-level annotated leaf instances. To our knowledge, this is the first instance segmentation dataset specifically designed for forestry leaves in open-field conditions. We propose LeafInst, a novel segmentation framework tailored for irregular and multi-scale leaf structures. The model integrates an Asymptotic Feature Pyramid Network (AFPN) for multi-scale perception, a Dynamic Asymmetric Spatial Perception (DASP) module for irregular shape modeling, and a dual-residual Dynamic Anomalous Regression Head (DARH) with Top-down Concatenation decoder Feature Fusion (TCFU) to improve detection and segmentation performance. On Poplar-leaf, LeafInst achieves 68.4 mAP, outperforming YOLOv11 by 7.1 percent and MaskDINO by 6.5 percent. On the public PhenoBench benchmark, it reaches 52.7 box mAP, exceeding MaskDINO by 3.4 percent. Additional experiments demonstrate strong generalization and practical utility for large-scale leaf phenotyping.
Paper Structure (27 sections, 21 equations, 17 figures, 9 tables)

This paper contains 27 sections, 21 equations, 17 figures, 9 tables.

Table of Contents

  1. Introduction
  2. Related Work
  3. Plant Phenotype Research Based on the Forestry Remote Sensing Technologies
  4. Instances Segmentation Models Based on the Deep Learning
  5. Methodology
  6. Model Neck
  7. Dynamic Anomalous Regression Head
  8. Dynamic Asymmetric Spatial Perception
  9. Top-down Concatenation - decoder Feature Fusion
  10. Loss Function
  11. Classification loss.
  12. Bounding box regression loss.
  13. Centerness loss.
  14. Mask loss.
  15. Experiment Settings
  16. ...and 12 more sections

Figures (17)

  • Figure 1: Challenges in forestry leaf phenotyping. (a) Scale changes, (b) Brightness changes, (c) Shape changes.
  • Figure 2: The workflow and downstream applications of the proposed framework LeafInst.
  • Figure 3: Overview of the main structure of the proposed LeafInst.
  • Figure 4: Adaptive spatial fusion operation (ASFF) built in AFPN.
  • Figure 5: The workflow of dynamic anomalous regression head.
  • ...and 12 more figures