A Cost-Effective Framework for Transforming Conventional Microscopes into Intelligent Diagnostic Platforms using Systematic Multi-Field Acquisition Protocol and Multimodal LLMs
DOI:
https://doi.org/10.59992/IJCI.2026.v5n5p3Keywords:
AI-Powered Microscopy, Urinalysis, Field-of-view Multimodal LLMs, Digital Pathology, Prompt EngineeringAbstract
Urine and stool cultures under the microscope are a main method of diagnostic work, but manual analysis is not only dependent on experience of the examiner but also suffers visual fatigue and subjectivity. Although, automated systems are available, their cost is so high that they are not often integrated in resource-limited environments. In this paper,we proposes a MicroScan AI, a low-cost solution that will be used to convert a standard optical microscope into a smart platform capable of diagnostic tasks. Morphological reasoning is done by connecting a digital microscopic camera with a 0.5x reduction lens with multimodal Large Language Models (LLM), namely ChatGPT-4o, Gemini, and Claude, and conducting the reasoning through encrypted API channels. A counting-based aggregation methodology was used to counteract a 60% field-of-view (FOV) hardware-induced field limit by a structured five-field acquisition scheme (F1–F5) scheme and Systematic Multi-Field for image pre-processing protocol . The validation of the experiment was done using 20 clinical specimens and 200 diagnostic parameters against expert human opinion. The system had an agreement rate of 97% with a Cohen Kappa coefficient of 0.94 which signifies an almost perfect agreement. Examination of the 3-percent discrepancy indicated that the error occurred in adjoining semi-quantitative grading frequencies and not resultant categorical detection errors. These results make MicroScan AI a scalable clinical decision support system that improves diagnostic consistency and decreases operator dependency without necessitating the complete replacement of hardware.
References
S. Haymond and C. McCudden, “Rise of the Machines: Artificial Intelligence and the Clinical Laboratory.,” J. Appl. Lab. Med., 2021. [Online]. Available: https://academic.oup.com/jalm/article/6/6/1640/6348035
[2] P. Moretti, R. Franchi, T. M. Poluzzi, and S. Paltrinieri, “Analytical variability and uncertainty in canine leukocyte ratios obtained with manual counts.,” Vet. Rec., 2022. [Online]. Available: https://bvajournals.onlinelibrary.wiley.com/doi/10.1002/vetr.1628
[3] H. Xie, Y. Jia, and S. Liu, “Integration of artificial intelligence in clinical laboratory medicine: Advancements and challenges,” Interdiscip. Med., 2024. [Online]. Available: https://onlinelibrary.wiley.com/doi/10.1002/INMD.20230056
[4] A. Rahman, C. L. Cheng, R. Salim, H. G. Di Shen, W. M. Gallagher, and J. Iqbal, “Digital pathology and AI: enhancing molecular diagnostics in low- and middle-income countries,” Expert Rev. Mol. Diagn., vol. 25, no. 12, pp. 859–867, 2025, doi: 10.1080/14737159.2025.2581687.
[5] A. Molani et al., “Advances in Portable Optical Microscopy Using Cloud Technologies and Artificial Intelligence for Medical Applications,” Sensors, vol. 24, no. 20, 2024, doi: 10.3390/s24206682.
[6] A. Mullan and A. Marsh, “Choosing the Best Camera for fluorescence Microscopy,” J. R. Microsc. Soc., pp. 16–22, 2019, doi: 10.22443/rms.inf.1.172.
[7] I. Kim, K. Kang, Y. Song, and T.-J. Kim, “Application of Artificial Intelligence in Pathology: Trends and Challenges,” Diagnostics, vol. 12, no. 11, 2022, doi: 10.3390/diagnostics12112794.
[8] R. Liu, W. Dai, T. Wu, M. Wang, S. Wan, and J. Liu, “AIMIC: Deep Learning for Microscopic Image Classification,” Comput. Methods Programs Biomed., vol. 226, p. 107162, 2022, doi: https://doi.org/10.1016/j.cmpb.2022.107162.
[9] Y. Rivenson, Z. Göröcs, H. Günaydin, Y. Zhang, H. Wang, and A. Ozcan, “Deep learning microscopy,” Optica, vol. 4, no. 11, pp. 1437–1443, 2017, doi: 10.1364/OPTICA.4.001437.
[10] J. von Bahr, A. Suutala, V. Diwan, A. Mårtensson, J. Lundin, and N. Linder, “AI-Supported Digital Microscopy Diagnostics in Primary Health Care Laboratories: Scoping Review,” J Med Internet Res, vol. 28, p. e78500, 2026, doi: 10.2196/78500.
[11] S. K. Mondal, M. Islam, and M. S. U. Yusuf, “An artificial intelligence-based approach for human parasite egg segmentation and classification,” Healthc. Anal., vol. 8, p. 100432, 2025, doi: https://doi.org/10.1016/j.health.2025.100432.
[12] K. M. Naing et al., “Automatic recognition of parasitic products in stool examination using object detection approach,” PeerJ Comput. Sci., vol. 8, p. e1065, 2022, doi: 10.7717/peerj-cs.1065.
[13] M. Li, Q. Gao, and T. Yu, “Kappa statistic considerations in evaluating inter-rater reliability between two raters: which, when and context matters,” pp. 1–5, 2023.
[14] M. Ji, G. Z. Genchev, H. Huang, T. Xu, H. Lu, and G. Yu, “Evaluation Framework for Successful Artificial Intelligence–Enabled Clinical Decision Support Systems: Mixed Methods Study,” J Med Internet Res, vol. 23, no. 6, p. e25929, 2021, doi: 10.2196/25929.
[15] J. Cadamuro et al., Clin. Chem. Lab. Med., vol. 63, no. 4, pp. 692–703, 2025, doi: 10.1515/cclm-2024-1016.
[16] H. Go, “Digital Pathology and Artificial Intelligence Applications in Pathology,” vol. 10, no. 2, pp. 76–82, 2022.
[17] M. Sonntagbauer, M. Haar, and S. Kluge, “[Artificial intelligence: How will ChatGPT and other AI applications change our everyday medical practice?].,” Med. Klin. Intensivmed. Notfmed., vol. 118, no. 5, pp. 366–371, Jun. 2023, doi: 10.1007/s00063-023-01019-6.
[18] R. Kaczmarczyk, T. Wilhelm, R. Martin, and J. Roos, “Evaluating multimodal AI in medical diagnostics,” npj Digit. Med., vol. 7, 2024, doi: 10.1038/s41746-024-01208-3.
[19] M. Harfouche et al., “Imaging across multiple spatial scales with the multi-camera array microscope,” Optica, vol. 10, no. 4, pp. 471–480, 2023, doi: 10.1364/OPTICA.478010.
[20] W.-C. Chen, C.-C. Chang, and Y. E. Lin, “Pulmonary Tuberculosis Diagnosis Using an Intelligent Microscopy Scanner and Image Recognition Model for Improved Acid-Fast Bacilli Detection in Smears,” Microorganisms, 2024. [Online]. Available: https://www.mdpi.com/2076-2607/12/8/1734
[21] E. Clarke et al., “Faster than light (microscopy): superiority of digital pathology over microscopy for assessment of immunohistochemistry,” J. Clin. Pathol., vol. 76, no. 5, pp. 333–338, 2023, doi: 10.1136/jclinpath-2021-207961.
[22] K. Basak, “Whole Slide Images in Artificial Intelligence Applications in Digital Pathology: Challenges and Pitfalls,” pp. 101–108, 2023, doi: 10.5146/tjpath.2023.01601.
[23] L. Budach et al., “The Effects of Data Quality on Machine Learning Performance,” 2022. [Online]. Available: https://www.semanticscholar.org/paper/b70226c67cfd7225a18f9c9efc61d1344584b507
[24] K. Nurzynska, M. Strzelecki, A. Piórkowski, and R. Obuchowicz, “AI in Medical Imaging and Image Processing,” J. Clin. Med., vol. 14, no. 12, 2025, doi: 10.3390/jcm14124153.
[25] C. Liu, D. Li, and P. Huang, “ISE-YOLO: Improved Squeeze-and-Excitation Attention Module based YOLO for Blood Cells Detection,” 2021 IEEE Int. Conf. Big Data (Big Data), 2021. [Online]. Available: https://doi.org/10.1109/bigdata52589.2021.9672069
[26] T. Islam, M. S. Hafiz, J. Jim, M. Kabir, and M. Ph.D., “A systematic review of deep learning data augmentation in medical imaging: Recent advances and future research directions,” Healthc. Anal., vol. 5, p. 100340, 2024, doi: 10.1016/j.health.2024.100340.
[27] B. D. Simon, K. B. Ozyoruk, D. G. Gelikman, S. A. Harmon, and B. Türkbey, “The future of multimodal artificial intelligence models for integrating imaging and clinical metadata: a narrative review.,” Diagn. Interv. Radiol., vol. 31, no. 4, pp. 303–312, Jul. 2025, doi: 10.4274/dir.2024.242631.
