Jove
Visualize
Contact Us
JoVE
x logofacebook logolinkedin logoyoutube logo
ABOUT JoVE
OverviewLeadershipBlogJoVE Help Center
AUTHORS
Publishing ProcessEditorial BoardScope & PoliciesPeer ReviewFAQSubmit
LIBRARIANS
TestimonialsSubscriptionsAccessResourcesLibrary Advisory BoardFAQ
RESEARCH
JoVE JournalMethods CollectionsJoVE Encyclopedia of ExperimentsArchive
EDUCATION
JoVE CoreJoVE BusinessJoVE Science EducationJoVE Lab ManualFaculty Resource CenterFaculty Site
Terms & Conditions of Use
Privacy Policy
Policies

Related Experiment Videos

PlantFormer: a precise plant disease segmentation network with interactive backbone and global-anisotropic context

Cuibiao Feng1,2, Yulong Fan3, Xi Chen3

  • 1Yiwu Industrial & Commercial College, Yiwu, China.

Frontiers in Plant Science
|July 11, 2026
PubMed
Summary

Related Concept Videos

You might also read

Related Articles

Articles linked to this work by shared authors, journal, and citation graph.

Sort by
Same author

[Separation and determination of furanocoumarins in shatian pomelo juice by HPLC-MS].

Se pu = Chinese journal of chromatography·2007
Same author

Effect of neuregulin-1 on histopathological and functional outcome after controlled cortical impact in mice.

Journal of neurotrauma·2007
Same author

Decreased expression of ING2 gene and its clinicopathological significance in hepatocellular carcinoma.

Cancer letters·2007
Same author

Compound Salvia droplet pill, a traditional Chinese medicine, for the treatment of unstable angina pectoris: a systematic review.

Medical science monitor : international medical journal of experimental and clinical research·2007
Same author

Ezrin silencing by small hairpin RNA reverses metastatic behaviors of human breast cancer cells.

Cancer letters·2007
Same author

The effect of nitric oxide on metal release from metallothionein-3: gradual unfolding of the protein.

Journal of biological inorganic chemistry : JBIC : a publication of the Society of Biological Inorganic Chemistry·2007

PlantFormer, a novel deep learning model, precisely segments plant diseases by addressing anisotropic spread and background noise. It achieves high accuracy on complex datasets, outperforming existing models for agricultural applications.

Area of Science:

  • Computer Vision
  • Plant Pathology
  • Machine Learning

Background:

  • Accurate plant disease segmentation is crucial for agriculture but challenged by lesion anisotropy, blurred boundaries, and background noise.
  • General-purpose models struggle with these domain-specific complexities in real-world field conditions.

Purpose of the Study:

  • To develop an end-to-end network, PlantFormer, specifically designed to overcome the limitations of current models in plant disease segmentation.
  • To improve the precision and robustness of automated plant disease identification in agricultural settings.

Main Methods:

  • PlantFormer utilizes an InteractSwin Backbone with Cross-Level Fusion (CLF) for preserving early pathological details.
  • A GlobalAnisotropic Context Aggregation (GACA) neck with strip pooling models directional disease spread.
Keywords:
context aggregationdeep learningplant disease segmentationsemantic segmentationswin transformer

Related Experiment Videos

  • A Semantic-Guided Fusion (SGF) decoder and a decoupled boundary-aware loss function refine segmentation and focus on disease margins.

    • Main Results:

    • PlantFormer achieved 41.78% mIoU on the challenging PlantSeg dataset and 93.54% mIoU on the structured NLB dataset.
    • The model demonstrated superior performance compared to DeepLabV3+ and Segformer in key metrics (mIoU, mAcc).

    Conclusions:

    • PlantFormer effectively addresses challenges in plant disease segmentation, offering a significant advancement for agricultural applications.
    • Future work will focus on enhancing performance in high-density, early-stage disease outbreak scenarios.