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 Concept Videos

You might also read

Related Articles

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

Sort by
Same author

Acid-treated ZIF-67-derived porous carbon based QuEChERS coupled with LC-MS/MS for sensitive determination of multi-pesticide residues.

Journal of chromatography. A·2026
Same author

Smartphone-Based Proactive Self-Screening for Ocular Surface Malignancies: A Nonrandomized Clinical Trial.

JAMA ophthalmology·2026
Same author

Application of a deep potential model for the structural evolution and phase stability of thorium.

The Journal of chemical physics·2026
Same author

Chemical composition imaging and class imbalance handling using a 2D-CNN with SMOTE and threshold moving for tobacco origin classification.

Talanta·2026
Same author

Study on the Mechanism of Buyang Huanwu Decoction in Treating Ischemic Stroke by Regulating the NLRP3/Caspase-1 Signaling Pathway.

Pharmaceuticals (Basel, Switzerland)·2026
Same author

IGF1/IGF-1R promotes hepatocellular carcinoma progression by activating the Akt/GSK-3β pathway.

PloS one·2026

Related Experiment Video

Updated: May 21, 2025

Influence of Hybrid Perovskite Fabrication Methods on Film Formation, Electronic Structure, and Solar Cell Performance
11:38

Influence of Hybrid Perovskite Fabrication Methods on Film Formation, Electronic Structure, and Solar Cell Performance

Published on: February 27, 2017

18.4K

Quantitative Analysis of Perovskite Morphologies Employing Deep Learning Framework Enables Accurate Solar Cell

Haixin Zhou1, Kuo Wang2,3, Cong Nie1

  • 1College of Railway Transportation, Hunan University of Technology, Zhuzhou, 412008, China.

Small (Weinheim an Der Bergstrasse, Germany)
|March 20, 2025
PubMed
Summary

A novel deep learning model, Self-UNet, accurately detects perovskite grain boundaries in SEM images. This enables precise prediction of solar cell efficiency, improving performance analysis.

Keywords:
deep learningdevice performance predictionedge detectiongrain boundaryperovskite solar cells

More Related Videos

Morphology Control for Fully Printable Organic–Inorganic Bulk-heterojunction Solar Cells Based on a Ti-alkoxide and Semiconducting Polymer
08:29

Morphology Control for Fully Printable Organic–Inorganic Bulk-heterojunction Solar Cells Based on a Ti-alkoxide and Semiconducting Polymer

Published on: January 10, 2017

9.0K
Flash Infrared Annealing for Perovskite Solar Cell Processing
05:15

Flash Infrared Annealing for Perovskite Solar Cell Processing

Published on: February 3, 2021

7.7K

Related Experiment Videos

Last Updated: May 21, 2025

Influence of Hybrid Perovskite Fabrication Methods on Film Formation, Electronic Structure, and Solar Cell Performance
11:38

Influence of Hybrid Perovskite Fabrication Methods on Film Formation, Electronic Structure, and Solar Cell Performance

Published on: February 27, 2017

18.4K
Morphology Control for Fully Printable Organic–Inorganic Bulk-heterojunction Solar Cells Based on a Ti-alkoxide and Semiconducting Polymer
08:29

Morphology Control for Fully Printable Organic–Inorganic Bulk-heterojunction Solar Cells Based on a Ti-alkoxide and Semiconducting Polymer

Published on: January 10, 2017

9.0K
Flash Infrared Annealing for Perovskite Solar Cell Processing
05:15

Flash Infrared Annealing for Perovskite Solar Cell Processing

Published on: February 3, 2021

7.7K

Area of Science:

  • Materials Science
  • Renewable Energy
  • Artificial Intelligence

Background:

  • Grain boundaries in perovskite solar cells are critical defect sites impacting performance.
  • Accurate quantification of grain morphology is essential for predicting solar cell efficiency.

Purpose of the Study:

  • To develop a deep learning model for precise perovskite grain boundary detection and morphological analysis.
  • To establish a correlation between grain morphology and perovskite solar cell performance.

Main Methods:

  • Development of the Self-UNet deep learning model for edge detection in scanning electron microscope (SEM) images.
  • Quantification of morphological features: grain boundary length (GBL), number of grains (NG), and average grain surface area (AGSA).
  • Integration of gradient boosted decision tree (GBDT) regression for predicting solar cell efficiency.

Main Results:

  • Self-UNet demonstrated superior edge detection accuracy (Dice: 91.22%, F1: 93.58%) compared to Canny and UNet.
  • The model accurately identified fine grains and distinguished true boundaries from surface grooves in low-quality SEM images.
  • GBDT regression achieved high accuracy in predicting solar cell efficiency with <10% relative error.

Conclusions:

  • Self-UNet provides a robust method for analyzing perovskite microstructure and its impact on device performance.
  • Accurate morphological quantification via Self-UNet enables reliable prediction of solar cell efficiency.
  • Grain boundary length (GBL) is validated as a key morphological feature for performance prediction.