Computer vision and deep learning-based prediction for inkjet-printed electrodes

Abstract

With the development of inkjet-printed electrodes, artificial intelligence-based quality control is essential for classifying inkjet-printed electrodes in a quality control environment. The quality of printed structures can be significantly affected by defects such as cracks, smudging, and misaligned deposits, which can degrade electrical performance and overall device reliability. Traditional quality control methods, including manual inspection and electrical testing, are time-consuming, subjective, and invasive, and they are unsuitable for high-throughput manufacturing environments. This work explores the application of computer vision and deep learning, specifically Convolutional Neural Networks (CNNs) and Feedforward Neural Networks, to automate defect detection and quality classification of inkjet-printed electrodes. To demonstrate the accessibility of deep learning techniques, Neural Architecture Search was implemented, showing the importance of automated model design in achieving high performance without extensive manual tuning or the need for expertise. The CNN models proved to be the most suitable approach for this image classification task, achieving a testing accuracy of 90.9% and a precision of 88.9% for a dataset of 2,406 electrode images containing both high-quality (1,020) and low-quality (1,386) prints.

Keywords

Inkjet printing / Electrodes / Defect detection / Deep learning / Computer vision / Convolutional Neural Networks / Feedforward neural networks / Neural architecture search

Acknowledgments

None.

Funding

This publication arises from research supported by Research Ireland under Grant Number 21/RC/10295_P2 and is co-funded by the European Regional Development Fund. This work is supported by I-Form.

Conflict of interest

Dermot Brabazon is one of the Associate Editors of the journal but was not involved in the editorial and peer-review process conducted for this paper, directly or indirectly. Separately, other authors declared that they have no known competing financial interests or personal relationships that could have influenced the work reported in this paper.

Author contributions

Conceptualization: Gareth Quinn, Anesu Nyabadza, Dermot Brabazon

Formal analysis: Gareth Quinn

Funding acquisition: Dermot Brabazon

Investigation: Gareth Quinn, Achu Titus

Methodology: Gareth Quinn

Supervision: Dermot Brabazon

Writing-original draft: Achu Titus

Writing-review & editing: Gareth Quinn, Sithara Sreenilayam, Éanna McCarthy, Anesu Nyabadza, Dermot Brabazon

Ethics approval and consent to participate

Not applicable.

Consent for publication

Not applicable.

Availability of data

The data supporting this article have been included in the Supplementary File.

References

[1] Vida J, Solak S, Shao Y, Homola T, List-Kratochvil E, Hermerschmidt F. Sintering of inkjet-printed silver nanoparticles by large-area atmospheric pressure nitrogen plasma. Appl Phys A. 2025; 131(2):91. doi: 10.1007/s00339-024-08206-y

[2] Du X, Wankhede SP, Prasad S, et al. A review of inkjet printing technology for personalized-healthcare wearable devices. J Mater Chem C. 2022; 10(38):14091-14115. doi: 10.1039/D2TC02511F

[3] Carey T, Cacovich S, Divitini G, et al. Fully inkjet-printed two-dimensional material field-effect heterojunctions for wearable and textile electronics. Nat Commun. 2017; 8(1):1202. doi: 10.1038/s41467-017-01210-2

[4] Rosati G, Cisotto G, Sili D, et al. Inkjet-printed fully customizable and low-cost electrodes matrix for gesture recognition. Sci Rep. 2021; 11(1):14938. doi: 10.1038/s41598-021-94526-5

[5] Yao C, Wang L, Wang Q, et al. Deep-learning-guided evaluation method for the high-volume preparation of flexible sensors based on inkjet printing. ACS Appl Interfaces Mater. 2024; 16(10):13326-13334. doi: 10.1021/acsami.4c00322

[6] Rossetti M, Srisomwat C, Urban M, et al. Unleashing inkjet-printed nanostructured electrodes and battery-free potentiostat for the DNA-based multiplexed detection of SARS-CoV-2 genes. Biosens Bioelectron. 2024; 250:116079. doi: 10.1016/j.bios.2024.116079

[7] Mei Y, Cannizzaro C, Park H, et al. Cell-compatible, multi-component protein arrays with subcellular feature resolution. Small. 2008; 4(10):1600. doi: 10.1002/smll.200800363

[8] Sarma Choudhury S, Katiyar N, Saha R, Bhattacharya S. Inkjet-printed flexible planar Zn-MnO₂battery on paper substrate. Sci Rep. 2024; 14(1):1597. doi: 10.1038/s41598-024-51871-5

[9] Kumar P, Ebbens S, Zhao X, et al. Inkjet printing of mammalian cells-theory and applications. Bioprinting. 2021; 23:e00157. doi: 10.1016/j.bprint.2021.e00157

[10] Bernasconi R, Invernizzi GP, Gallo Stampino E, Gotti R, Gatti D, Magagnin L. Printing MEMS: Application of inkjet techniques to the manufacturing of inertial accelerometers. Micromachines (Basel). 2023; 14(11):2082. doi: 10.3390/mi14112082

[11] Chen S, He Z, Choi S, Novosselov IV. Characterization of inkjet-printed digital microfluidics devices. Sensors (Basel). 2021; 21(9):3064. doi: 10.3390/s21093064

[12] Tian L, Liu J, Chen X, Branicio PS, Lei Q. Mechanisms and strategies to achieve stability in inkjet printed 2D materials electronics. Adv Electron Mater. 2025; 11(3):2400143. doi: 10.1002/aelm.202400143

[13] Yunker PJ, Still T, Lohr MA, Yodh AG. Suppression of the coffee-ring effect by shape-dependent capillary interactions. Nature. 2011; 476(7360):308-311. doi: 10.1038/nature10344

[14] Soltman D, Subramanian V. Inkjet-printed line morphologies and temperature control of the coffee ring effect. Langmuir. 2008; 24(5):2224-2231. doi: 10.1021/la7026847

[15] Sowade E, Ramon E, Mitra KY, et al. All-inkjet-printed thin-film transistors: Manufacturing process reliability by root cause analysis. Sci Rep. 2016; 6(1):33490. doi: 10.1038/srep33490

[16] Chen Z, Gengenbach U, Koker L, et al. Systematic investigation of novel, controlled low‐temperature sintering processes for inkjet printed silver nanoparticle ink. Small. 2024; 20(21):2306865. doi: 10.1002/smll.202306865

[17] Kwon KS, Jo JY. Towards zero-defect inkjet printing via piezo self-sensing signals. Sens Actuators A Phys. 2025; 393:116755. doi: 10.1016/j.sna.2025.116755

[18] Kwon KS. Methods for detecting air bubble in Piezo inkjet dispensers. Sens Actuators A Phys. 2009; 153(1):50-56. doi: 10.1016/j.sna.2009.04.024

[19] Jeurissen R, De Jong J, Reinten H, et al. Effect of an entrained air bubble on the acoustics of an ink channel. J Acoust Soc Am. 2008; 123(5):2496-2505. doi: 10.1121/1.2835624

[20] Law KNC, Yu M, Zhang L, et al. Enhancing Printed Circuit Board Defect Detection through Ensemble Learning. New Jersey: IEEE; 2024. p. 37-44. doi: 10.1109/FITYR63263.2024.00013

[21] Novikovs A, Tsebriienko T, Trausa A, et al. Novel terpineol-based silver nanoparticle ink with high stability for inkjet printing. Nanomaterials (Basel). 2025; 15(13):955. doi: 10.3390/nano15130955

[22] Beedasy V, Smith PJ. Printed electronics as prepared by inkjet printing. Materials (Basel). 2020; 13(3):704. doi: 10.3390/ma13030704

[23] Amini A, Gan TH. A computer vision-based quality assessment technique for R2R printed silver conductors on flexible plastic substrates. Appl Sci. 2023; 13(2):1084. doi: 10.3390/app13021084

[24] Kim J, Jung Y, Parajuli S, et al. Quantifying printing quality for printed electrodes via deep learning and spatial association: Empowering process optimization. Adv Intell Syst. 2025; 7:2500178. doi: 10.1002/aisy.202500178

[25] Ogunsanya M, Isichei J, Parupelli SK, Desai S, Cai Y. In-situ droplet monitoring of inkjet 3D printing process using image analysis and machine learning models. Proced Manuf. 2021; 53:427-434. doi: 10.1016/j.promfg.2021.06.045

[26] Huang J, Segura LJ, Wang T, Zhao G, Sun H, Zhou C. Unsupervised learning for the droplet evolution prediction and process dynamics understanding in inkjet printing. Addit Manuf. 2020; 35:101197. doi: 10.1016/j.addma.2020.101197

[27] Nandipati M, Ogunsanya M, Desai S, et al. Predictive models for 3D inkjet material printer using automated image analysis and machine learning algorithms. Manuf Lett. 2024; 41:810-821. doi: 10.1016/j.mfglet.2024.09.101

[28] Polomoshnov M, Reichert KM, Rettenberger L, et al. Image-based identification of optical quality and functional properties in inkjet-printed electronics using machine learning. J Intell Manuf. 2025; 36(4):2709-2726. doi: 10.1007/s10845-024-02385-4

[29] Carou-Senra P, Ong JJ, Castro BM, et al. Predicting pharmaceutical inkjet printing outcomes using machine learning. Int J Pharm X. 2023; 5:100181. doi: 10.1016/j.ijpx.2023.100181

[30] Sahu A, Aaen PH, Damacharla P, et al. An Automated Machine Learning Approach to Inkjet Printed Component Analysis: A Step Toward Smart Additive Manufacturing. New York: IEEE; 2024. p. 1-6. doi: 10.1109/WMCS62019.2024.10618993

[31] Lall P, Soni V, Kulkarni S, Miller S, et al. Comparison of Machine Learning Approaches for Correlating Print Process Parameters to Realized Physical and Electrical Characteristics of Printed Electronics Using Inkjet Platform. New York: American Society of Mechanical Engineers; 2023. p. V001T03A012. doi: 10.1115/IPACK2023-112056

[32] Queralto A, Pacheco A, Jimenez N, et al. Defining inkjet printing conditions of superconducting cuprate films through machine learning. J Mater Chem C. 2022; 10(17):6885-6895. doi: 10.1039/d1tc05913k

[33] Kim S, Cho M, Jung S. The design of an inkjet drive waveform using machine learning. Sci Rep. 2022; 12(1):4841. doi: 10.1038/s41598-022-08784-y

[34] Kim S, Wenger R, Bürgy O, et al. Predicting inkjet jetting behavior for viscoelastic inks using machine learning. Flex Print Electron. 2023; 8(3):035007. doi: 10.1088/2058-8585/acee94

[35] Jiang L, Wolf R, Alharbi K, Qin H, et al. In situ monitoring and recognition of printing quality in electrohydrodynamic inkjet printing via machine learning. J Manuf Sci Eng. 2024; 146(11):110901.

[36] Brishty FP, Urner R, Grau G, et al. Machine learning based data driven inkjet printed electronics: Jetting prediction for novel inks. Flexible and Printed Electronics. 2022; 7(1):015009. doi: 10.1088/2058-8585/ac5a39

[37] Kim S, Cho M, Jung S, et al. Reinforcement learning-based dynamic optimization of driving waveforms for inkjet printing of viscoelastic fluids. Langmuir. 2025; 41(17):10831-10840. doi: 10.1021/acs.langmuir.4c05141

[38] Phung TH, Park SH, Kim I, Lee TM, Kwon KS. Machine learning approach to monitor inkjet jetting status based on the piezo self-sensing. Sci Rep. 2023; 13(1):18089. doi: 10.1038/s41598-023-45445-0

[39] Li M, Yin S, Liu Z, Zhang H. Machine learning enables electrical resistivity modeling of printed lines in aerosol jet 3D printing. Sci Rep. 2024; 14(1):14614. doi: 10.1038/s41598-024-65693-y

[40] Shirsavar MA, Taghavimehr M, Ouedraogo LJ, et al. Machine learning-assisted E-jet printing for manufacturing of organic flexible electronics. Biosens Bioelectron. 2022; 212:114418. doi: 10.1016/j.bios.2022.114418

[41] Nyabadza A, Titus A, McCarthy É, et al. Fabrication and inkjet printing of manganese oxide electrodes for energy storage. Chem Eng J Adv. 2025; 22:100761. doi: 10.1016/j.ceja.2025.100761