Research Projects

Research questions, illustrated method overviews, and associated publications. The diagrams show conceptual workflows rather than recordings or measured results.

For a broader overview, visit Research Activities. Publication titles and DOI links below are sourced from the site's curated bibliography.

Plant Bioelectrical Potential Acquisition for Edge Sensing

Active

Role: PhD Student Institution: Kanazawa University

Research on plant bioelectrical-potential acquisition, signal analysis, and sensing architectures for environmental and edge-sensing applications.

Research question
Which acquisition and signal-processing choices support trustworthy plant bioelectric measurements?
Approach
Review electrode placement, recording hardware, measurement environments, and edge-sensing considerations reported in the literature.
Method overviewAcquisition and signal quality
  1. Plant signalsBioelectric potential measurements
  2. RecordingElectrodes and acquisition setup
  3. Signal qualityNoise, drift and recording conditions
  4. AnalysisEvaluate usable recordings and methods
Conceptual research workflow — not a signal recording, experimental result, equipment photograph or a claim of deployed functionality.

Related publication

Collaborators: Hidetaka Nambo

Plant Bioelectrical SignalsSignal ProcessingEnvironmental SensingEdge Sensing
Link to this project →

Transformer-Based Models for Plant Bioelectric Signal Classification

Active

Role: PhD Student Institution: Kanazawa University

Research on machine-learning approaches for classifying plant bioelectric signals, including Transformer-based models for noisy biological time-series data.

Research question
How can noisy plant bioelectric time-series be classified and evaluated reliably?
Approach
Study sequence classifiers, including Transformer-based models, with attention to preprocessing and validation.
Method overviewTime-series classification
  1. Signal sequencesPlant bioelectric time-series
  2. PreprocessingSignal preparation choices
  3. Sequence modelsCompare model approaches
  4. EvaluationValidation and reproducibility
Conceptual research workflow — not a signal recording, experimental result, equipment photograph or a claim of deployed functionality.

Published research evidence · 2026

Evidence base of the BEP Transformer scoping review

How the 182 included studies are categorized in the published review. This is a chart of reviewed literature, not plant-signal measurements or classifier performance.

Direct BEP and plant-domain studies Plant bioelectric / domain evidence
18
Transformer / time-series bottleneck studies Model limitation evidence
13
Analogical adaptation-source studies Adjacent-domain methods; not BEP validation
151
Original visualization of counts reported in the scoping review abstract (18 + 13 + 151 = 182). The 151 analogical-source studies do not demonstrate BEP-specific effectiveness. Article details Publisher DOI ↗

Related publications

Collaborators: Hidetaka Nambo

PyTorchTransformersTime-Series AnalysisDeep Learning
Link to this project →

GAN-Augmented Deep Learning for Malaria Classification

Completed

Role: Co-Author Institution: Universitas Amikom Purwokerto

Applied computer-vision research combining GAN-based data augmentation and transfer learning for malaria classification from microscopy images.

Research question
How can synthetic augmentation and transfer learning support microscopy-based malaria classification?
Approach
Investigate GAN-augmented image data and pretrained computer-vision models as described in the related publication.
Method overviewMicroscopy image analysis
  1. Image dataMicroscopy inputs
  2. AugmentationGAN-based training support
  3. Transfer learningPretrained visual representations
  4. ClassificationEvaluate image classification
Conceptual research workflow — not a signal recording, experimental result, equipment photograph or a claim of deployed functionality.

Related publication

Collaborators: Chyntia Raras Ajeng Widiawati, Hikmalul A'la Syahrizaldy, Karina Puspita Prameswari

Deep LearningGANsTransfer LearningComputer Vision
Link to this project →

GAN-Enhanced RBF Networks for Landslide Susceptibility Mapping

Completed

Role: Co-Author Institution: Universitas Amikom Purwokerto

Applied machine-learning research combining generative augmentation and radial-basis-function networks for landslide susceptibility mapping.

Research question
How can generative augmentation and RBF-based models contribute to landslide susceptibility mapping?
Approach
Examine GAN-based augmentation, geospatial predictors, and radial-basis-function modeling in the associated study.
Method overviewGeospatial modeling
  1. Spatial inputsGeographic predictors
  2. AugmentationGenerative modeling
  3. RBF networkSusceptibility estimation
  4. MappingSpatial model output
Conceptual research workflow — not a signal recording, experimental result, equipment photograph or a claim of deployed functionality.

Related publication

Collaborators: Chyntia Raras Ajeng Widiawati, Ika Maulita, Yuli Purwati

RBF NetworksGIS DataGenerative ModelingLandslide Susceptibility
Link to this project →