350 rub
Journal Biomedical Radioelectronics №3 for 2025 г.
Article in number:
Detection of key points of electronic planograms by means of artificial intelligence
Type of article: scientific article
DOI: https://doi.org/10.18127/j15604136-202503-05
UDC: 004.89
Keywords: Among the diseases of the feet flat feet are the most common cause of disorders of the musculoskeletal system. Flat feet are characterized by flattening of the arches of the foot decreased cushioning function changes in the biomechanics of walking and pain syndromes. For the diagnosis of flat feet computer plantography is widely used which is why an urgent task is to detect key points on electronic plantograms in order to diagnose the condition of the foot in this disease. The ef-fectiveness of this process can be improved by reducing its duration and objectification using artificial intelligence tech-nology. To develop and evaluate a model for detecting key points in computer plantography images using artificial intelligence. The study used a data set containing 1000 images of computer plantography of different patients allowing for a morpho-functional assessment of the foot and calculating deformity indicators. Based on the results of training the yolo11x-pose model (a model for determining the location of certain points in the im-age) the dynamics of changes in the values of the loss function and changes in the quality metrics of the model as it is trained were obtained. The detection accuracy was 0.9951 for objects and 0.9952 for points the detection completeness was 1.0 for both tasks. These results demonstrate the model's ability to correctly recognize target elements with minimal errors. At the same time mAP50 values were at the level of 0.995 and mAP50-95 were 0.9477 for object detection and 0.9647 for key point detection tasks. The detection of key points in computer plantography images of the artificial intelligence model makes it possible to in-crease the efficiency of detecting key points on electronic plantograms in order to diagnose the condition of the feet.
Authors:

V.V. Mikhailishin1, L.M. Smirnova2, V.A. Khazov3

1, 2 Federal Scientific and Educational Centre of Medial and Social Expertise and Rehabilitation n. a. G.A. Albreht (Saint-Petersburg, Russia)
2, 3 St. Petersburg State Electrotechnical University “LETI” (St. Petersburg, Russia)

Abstract:

Among the diseases of the feet, flat feet are the most common cause of disorders of the musculoskeletal system. Flat feet are characterized by flattening of the arches of the foot, decreased cushioning function, changes in the biomechanics of walking and pain syndromes. For the diagnosis of flat feet, computer plantography is widely used, which is why an urgent task is to detect key points on electronic plantograms in order to diagnose the condition of the foot in this disease. The effectiveness of this process can be improved by reducing its duration and objectification using artificial intelligence technology.

To develop and evaluate a model for detecting key points in computer plantography images using artificial intelligence.
The study used a data set containing 1000 images of computer plantography of different patients, allowing for a morpho-functional assessment of the foot and calculating deformity indicators.

Based on the results of training the yolo11x-pose model (a model for determining the location of certain points in the image), the dynamics of changes in the values of the loss function and changes in the quality metrics of the model as it is trained were obtained. The detection accuracy was 0.9951 for objects and 0.9952 for points, the detection completeness was 1.0 for both tasks. These results demonstrate the model's ability to correctly recognize target elements with minimal errors. At the same time, mAP50 values were at the level of 0.995, and mAP50-95 were 0.9477 for object detection and 0.9647 for key point detection tasks.
The detection of key points in computer plantography images of the artificial intelligence model makes it possible to increase the efficiency of detecting key points on electronic plantograms in order to diagnose the condition of the feet.

Pages: 42-47
For citation

Mikhailishin V.V., Smirnova L.M., Khazov V.A. Detection of key points of electronic planograms by means of artificial intelligence. Biomedicine Radioengineering. 2025. V. 28. № 3. P. 42–47. DOI: https:// doi.org/10.18127/j15604136-202503-05 (In Russian)

References
  1. Evans A. M., Rome K. A review of the evidence for non-surgical interventions for flexible pediatric flat feet. Eur J Phys Rehabil Med. 2011. V. 47. № 1. P. 1–21.
  2. Salinas-Torres V.M., Salinas-Torres R.A., Carranza-García L.E., Herrera-Orozco J., Tristán-Rodríguez J.L. Prevalence and Clinical Factors Associated With Pes Planus Among Children and Adults: A Population-Based Synthesis and Systematic Review. J. Foot Ankle Surg. 2023. Sep.-Oct. V. 62(5). P. 899–903. DOI: 10.1053/j.jfas.2023.05.007. Epub 2023 Jun 5. PMID: 37286098.
  3. Vedenina A.S., Tkachuk I.V., Smirnova L.M. i dr. Skrining funkcional'nyh narushenij stop s pomoshch'yu komp'yuternoj plantografii i podometrii. Medicinskaya tekhnika. 2014. № 2(284). S. 21–24 (In Russian).
  4. Eliseeva O.G., Gavrikov K.V., Krayushkin A.I. i dr. Ispol'zovanie metoda komp'yuternoj plantografii dlya issledovaniya morfofunkcional'nogo sostoyaniya stopy u detej. Byulleten' Volgogradskogo nauchnogo centra Rossijskoj akademii medicinskih nauk i Administracii Volgogradskoj oblasti. 2006. № 2. S. 18–18a (In Russian).
  5. Patent № 2253363 (RF). Sposob diagnostiki sostoyaniya otdelov stopy. K.V. Gavrikov, I.A. Pleshakov, S.I. Kaluzhskij, A.I. Perepelkin, N.V. Andreev. 2005. Byul. № 16 (In Russian).
  6. Patent № 2492803 (RF). Sposob diagnostiki ressornoj i opornoj funkcij stopy sportsmena. A.A. Rudenko, N.I. Ivanova. 2013 (In Russian).
  7. Perepelkin A.I., Krayushkin A.I., Atroshchenko E.S. Issledovanie prodol'nogo svoda stopy u detej metodom komp'yuternoj plantografii. Byulleten' medicinskih internet-konferencij. 2015. T. 5. № 7. S. 1053–1057 (In Russian).
  8. Castiglioni I. et al. AI applications to medical images: From machine learning to deep learning. Physica medica. 2021. V. 83. P. 9–24.
  9. Obuchowicz R., Strzelecki M., Piórkowski A. Clinical Applications of Artificial Intelligence in Medical Imaging and Image Processing–A Review. Cancers. 2024. V. 16. № 10. P. 1870.
  10. Tang X. The role of artificial intelligence in medical imaging research. BJR| Open. 2019. V. 2. № 1. P. 20190031.
  11. Mihajlishin V.V. Smirnova L.M., Cherkashin S.O. Cifrovaya obrabotka elektronnyh plantogramm s primeneniem tekhnologij iskusstvennogo intellekta kak etap avtomatizacii plantograficheskih issledovanij. Cifrovaya obrabotka signalov. 2024. № 3. S. 19–24 (In Russian).
  12. Ioffe R.Ya., Smirnova L.M., Belyanin O.L. Kompleksnaya ocenka sostoyaniya stopy na sisteme «Skan». Vestnik vserossijskoj gil'dii protezistov-ortopedov. 2004. № 2. S. 36–40 (In Russian).
  13. Ragab M. G. et al. A Comprehensive Systematic Review of YOLO for Medical Object Detection (2018 to 2023). IEEE Access. 2024.
Date of receipt: 05.02.2025
Approved after review: 16.02.2025
Accepted for publication: 29.05.2025