Skip to content
DraconiticResearch
Search Sources API
194,161 studies31 sourcesUpdated 6 min ago
Peer-reviewedCC BY 4.0PLOS One

MambaSeed: A lightweight Mamba-based hybrid vision network for tomato leaf disease recognition

  • XueShen Wang1
  • Heng Zhang2
  • Yayong Chen3
  • YunLong Li4
  • YuJie Qin
  • ZhongYu Lv
  • XiangJuan Liu
4 affiliations
  1. 1College of Computer and Control Engineering, Qiqihar University, Qiqihar, Heilongjiang, China
  2. 2State Key Laboratory of Water Engineering Ecology and Environment in Arid Area, Xi’an University of Technology, Xi’an, China
  3. 3School of Water Resources and Hydro-electric Engineering of XUT, Xi’an University of Technology, Xi’an, China
  4. 4Heilongjiang Key Laboratory of Big Data Network Security Detection and Analysis, Qiqihar University, Qiqihar, Heilongjiang, China
Open at sourcePDFDOICite

Abstract

Crop diseases threaten agricultural productivity, creating a need for recognition models that combine accuracy, robustness, and computational efficiency. Existing CNN-, Transformer-, and state-space-based approaches can weaken fine-grained lesion edges and textures when emphasizing global context. We therefore propose MambaSeed, a lightweight hybrid network built on MambaVisionMixer and equipped with the Local Continuous Branch Injection mechanism, hereafter LCBI. This mechanism repeatedly introduces local feature compensation during hierarchical modeling, preserving high-frequency cues while maintaining global dependency modeling. On the Tomato, Corn, and Potato datasets, MambaSeed achieved accuracies of 98.97%, 98.06%, and 95.87%, respectively, with 1.13M parameters and 0.240G MACs. In the batch-one FP32 GPU benchmark, its latency was 3.39 ms/image on an NVIDIA RTX 4090. Without a hardware accelerator, the median CPU latency was 25.35 ms/image with one thread and 15.61 ms/image with four threads. Ablation, cross-architecture frequency analysis, and controlled perturbation tests further showed that LCBI strengthens high-frequency feature retention and maintains stable recognition under visual degradation. Together, these results show that MambaSeed provides a compact balance of accuracy, efficiency, and robustness across both accelerator-based and CPU-only inference settings.

Related

  1. Lightweight frequency-aware hybrid vision transformer for fine-grained plant disease recognition
    Kun Li, Li Dong, Jing Qian +3
    Frontiers in Plant SciencePeer-reviewedCC BY 4.0Full textPlant biologyCrossrefPMC
  2. Citrus leaf pathology multi-class image dataset and a hybrid multi-branch attention fusion network for lemon leaf disease classification
    Chayan Mondal Abir, Subir Biswas, Prabodh Kumar Sahoo +1
    Scientific ReportsPeer-reviewedCC BY 4.0Full textPlant biologyCrossref
  3. Probing EEG Foundation Model Attention Maps for Lightweight Interpretable Emotion Recognition
    L. N. Arvanitopoulos, A. Dragomir, C. A. Fidas +1
    bioRxivPreprintCC BYFull textNeurosciencebioRxiv
  4. Soy-RepViT: a lightweight deep learning framework for fine-grained recognition of soybean growth habits
    Dongbo Liu, Daohan Cui, Hang Liu +5
    Frontiers in plant sciencePeer-reviewedCC BYFull textPlant biologyEurope PMCCrossref
  5. Lightweight CNN and transformer detectors for real-time single-fruit date-variety recognition on edge GPU hardware
    Adnan Miski, Moayad Alshangiti
    PLOS OnePeer-reviewedCC BY 4.0Full textGalaxiesNeurosciencePLOSPMC

Cite

BibTeXRIS
BibTeX
@article{wang2026mambaseed,
  title = {{MambaSeed: A lightweight Mamba-based hybrid vision network for tomato leaf disease recognition}},
  author = {XueShen Wang and Heng Zhang and Yayong Chen and YunLong Li and YuJie Qin and ZhongYu Lv and XiangJuan Liu},
  journal = {PLOS One},
  year = {2026},
  month = oct,
  doi = {10.1371/journal.pone.0359350},
  url = {https://doi.org/10.1371/journal.pone.0359350},
  note = {CC BY 4.0},
}
RIS
TY  - JOUR
TI  - MambaSeed: A lightweight Mamba-based hybrid vision network for tomato leaf disease recognition
AU  - XueShen Wang
AU  - Heng Zhang
AU  - Yayong Chen
AU  - YunLong Li
AU  - YuJie Qin
AU  - ZhongYu Lv
AU  - XiangJuan Liu
PY  - 2026
DA  - 2026/10/15
JO  - PLOS One
DO  - 10.1371/journal.pone.0359350
UR  - https://doi.org/10.1371/journal.pone.0359350
AB  - Crop diseases threaten agricultural productivity, creating a need for recognition models that combine accuracy, robustness, and computational efficiency. Existing CNN-, Transformer-, and state-space-based approaches can weaken fine-grained lesion edges and textures when emphasizing global context. We therefore propose MambaSeed, a lightweight hybrid network built on MambaVisionMixer and equipped with the Local Continuous Branch Injection mechanism, hereafter LCBI. This mechanism repeatedly introduces local feature compensation during hierarchical modeling, preserving high-frequency cues while maintaining global dependency modeling. On the Tomato, Corn, and Potato datasets, MambaSeed achieved accuracies of 98.97%, 98.06%, and 95.87%, respectively, with 1.13M parameters and 0.240G MACs. In the batch-one FP32 GPU benchmark, its latency was 3.39 ms/image on an NVIDIA RTX 4090. Without a hardware accelerator, the median CPU latency was 25.35 ms/image with one thread and 15.61 ms/image with four threads. Ablation, cross-architecture frequency analysis, and controlled perturbation tests further showed that LCBI strengthens high-frequency feature retention and maintains stable recognition under visual degradation. Together, these results show that MambaSeed provides a compact balance of accuracy, efficiency, and robustness across both accelerator-based and CPU-only inference settings.
SN  - 1932-6203
ER  -