Introduction

The measurement of foot contact area while standing is critical for assessing the presence and severity of conditions such as pes planus (flatfoot) and pes cavus (high arch), which affect pressure distribution across the plantar surface [1]. In individuals with pes planus, the medial longitudinal arch is reduced or absent, resulting in a larger ground contact area [2]. Conversely, individuals with pes cavus have an exaggerated arch, which leads to a smaller foot contact area [3]. Accurate measurement of foot contact area allows clinicians to indirectly evaluate the arch height and identify abnormal foot postures [2,3]. Anomalies such as pes planus and pes cavus disrupt this distribution, potentially causing functional problems or foot pain [4]. Precise foot contact area measurement offers an objective quantification of the severity of these conditions, enabling precise diagnosis and individualized treatment strategies.

Foot contact area measurement is typically performed using a podoscope, which consists of a transparent platform illuminated from below, allowing visualization of the plantar surface [5,6]. This helps physicians analyze the pressure distribution and foot contact area, identify conditions such as pes planus and pes cavus, and develop appropriate treatment plans [5,6]. While a podoscope provides a visual representation to intuitively assess the presence and degree of these conditions, no method currently exists to quantitatively measure and analyze foot contact area, making it challenging to objectively diagnose them or to compare changes in contact distribution during follow-up assessments. Therefore, developing a program that can automatically calculate foot contact area would enhance diagnostic accuracy and facilitate effective treatment planning. Quantitative assessment tools would enable clinicians to establish baseline measurements, monitor changes over time, and evaluate the effectiveness of therapeutic interventions.

ChatGPT-4, an advanced language model developed by OpenAI, is known for its ability to understand and generate humanlike text based on extensive training data [7,8]. This advanced technology has been applied across various fields, including healthcare, where it aids in data analysis, clinical decision-making, and the development of automated tools [7,8]. In addition, ChatGPT-4 can assist in coding and program development and provide valuable support throughout the programming process.

This study aimed to develop a program that automatically measures foot contact area using a podoscope, leveraging the capabilities of ChatGPT-4, and to present the detailed development process of the program.

Case

Ethics statement: This study was approved by the Institutional Review Board (IRB) of Yeungnam University Hospital (IRB No: 2024-10-030). Institutional approval was obtained for the publication of the case details. Written informed consent was obtained from the volunteer for the publication of this case report and accompanying images.

We aimed to develop a program that automatically measures foot contact area using ChatGPT-4. Initially, a 53-year-old female volunteer was instructed to stand on a podoscope and images of both feet were captured. The initial images contained unnecessary background elements that were removed, leaving only the foot shape for analysis. The edited image and prompt were input into ChatGPT-4 to achieve the desired outcomes. In the prompt (Fig. 1), the specific requirements for recognizing the shape of the foot with a blue outline, detecting the green (fluorescent) areas with a red outline, and calculating the respective areas and their ratios are clearly outlined. ChatGPT-4 generated Python code that met these criteria. The code created by ChatGPT-4 (Fig. 2) performs image processing tasks using OpenCV involving the following key steps:

1. The images were loaded and converted to grayscale to identify the shape of the foot.

2. Binarization was applied to separate the background from the foot and detect the contour of the foot, which is outlined in blue.

3. The image was converted to the hue/saturation/value color space to detect green (fluorescent) areas, which are marked in red.

4. The area of the foot and green (fluorescent) region were calculated, and the ratio of the two regions was determined.

5. Finally, the results were visualized to clearly present the outline of the foot and green contact area.

Each step of the code was effectively executed, thereby enhancing the reliability of the analysis with clear visual outputs. The processed image (Fig. 3) displays the outline of the foot, marked in blue, and the green (fluorescent) contact area, marked in red. Additionally, individual masks for the foot and green area were generated for calculation. The total area of the foot and contact area were recorded in pixels (total foot area, 1,091,381.00 pixels; ground contact area, 604,252.50 pixels). The ratio of the contact area to the total foot area was calculated as 55.37%.

Discussion

In this study, we utilized ChatGPT-4 to develop a program that identifies the entire foot and regions in contact with the ground when the foot is placed on a podoscope, calculates their respective areas, and determines the ratio between the two areas. This program offers an objective approach to measuring the severity of pes planus and pes cavus, providing valuable information for both physicians and patients during follow-up assessments. Conventional methods of visual inspection lack the precision necessary to accurately diagnose and monitor foot posture abnormalities. By leveraging the capabilities of ChatGPT-4 and image-processing techniques, this program enhances diagnostic accuracy.

Our findings demonstrate that the program effectively visualizes foot contours and contact areas. Automatic calculation of areas and their ratios offers a quantitative basis for clinical evaluation, making it particularly valuable for tracking changes over time. This quantitative assessment aids in diagnosing flatfoot and high-arch conditions and facilitates the evaluation of treatment efficacy.

By integrating advanced image processing algorithms, particularly through OpenCV, the program highlights the potential of machine learning and artificial intelligence in clinical settings. Its ability to recognize and delineate different foot regions provides clinicians with actionable insights that were previously difficult to quantify. Furthermore, the visual representation of results enhances the ability of clinicians to communicate findings to patients and promotes better understanding and engagement in treatment plans.

Our program, developed with the assistance of ChatGPT-4, does not automatically diagnose pes planus or pes cavus and was not intended to establish cutoff values. To determine the cutoff values for diagnosing these conditions, it is necessary to recruit many subjects and collect sufficient data. Our study serves as a pilot study, demonstrating that it is possible to easily create a program for automatically calculating foot contact areas using ChatGPT-4. We believe that our program can be used as a supplementary tool to assist in the diagnosis of pes planus and pes cavus. Moreover, because the current study involved a single case, further validation with a larger sample size is essential to establish the reliability and generalizability of the program across diverse populations. In addition, incorporating feedback from clinicians during the development process can enhance the user experience and clinical applicability of the program. In the future, a comparative analysis between this automated foot contact area measurement program for podoscopes, developed with the assistance of ChatGPT-4, and the pressure sensor-based foot scan system is necessary to demonstrate the accuracy of our developed program.

In conclusion, the program developed in this study represents a promising step toward automating foot contact area measurements. By providing objective and quantitative data, it has the potential to revolutionize the assessment and management of foot-related conditions, leading to improved patient outcomes and personalized treatment strategies. ChatGPT-4 can automatically code the program that physicians can easily adapt to meet specific clinical requirements. As the coding abilities of ChatGPT-4 continue to advance, physicians are expected to experience an increased ability to develop customized programs for their clinical practice, enhancing patient care and clinical workflow.

Article information

Conflicts of interest

Min Cheol Chang has been an associate editor of Journal of Yeungnam Medical Science since 2021. He was not involved in the review process of this manuscript. There are no other conflicts of interest to declare.

Funding

This work was supported by the National Research Foundation of Korea (NRF) grant funded by the Korea government (MSIT) (No. 00219725).

Fig. 1.

Prompt provided to ChatGPT-4 to recognize the foot shape, draw a blue outline around the perimeter, detect the green (fluorescent) area with a red outline, and calculate both areas and their ratio.

Fig. 2.

Python code generated by ChatGPT-4 utilizing OpenCV for image processing.

Fig. 3.

Visual output after executing the code shows the blue outline of the foot and the red outline of the green (fluorescent) area. Additionally, masks for the foot and contact area are displayed, along with pixel counts for the entire foot area and the ground contact area. The entire foot area measures 1,091,381.00 pixels, with a contact area of 604,252.50 pixels. The ratio of the ground contact area to the entire foot area is automatically calculated as 55.37%.

References

• 1. Buldt AK, Forghany S, Landorf KB, Levinger P, Murley GS, Menz HB. Foot posture is associated with plantar pressure during gait: a comparison of normal, planus and cavus feet. Gait Posture 2018;62:235–40.ArticlePubMed
• 2. Aenumulapalli A, Kulkarni MM, Gandotra AR. Prevalence of flexible flat foot in adults: a cross-sectional study. J Clin Diagn Re 2017;11:AC17–20.ArticlePubMedPMC
• 3. Maynou C, Szymanski C, Thiounn A. The adult cavus foot. EFORT Open Rev 2017;2:221–9.ArticlePubMedPMCPDF
• 4. Riskowski JL, Dufour AB, Hagedorn TJ, Hillstrom HJ, Casey VA, Hannan MT. Associations of foot posture and function to lower extremity pain: results from a population-based foot study. Arthritis Care Res (Hoboken) 2013;65:1804–12.ArticlePubMedPMC
• 5. Halabchi F, Mazaheri R, Mirshahi M, Abbasian L. Pediatric flexible flatfoot; clinical aspects and algorithmic approach. Iran J Pediatr 2013;23:247–60.PubMedPMC
• 6. Rdzanek J, Świątek J, Wychowański M. Correlation between foot arch parameters and body balance performance. Ortop Traumatol Rehabil 2022;24:263–72.ArticlePubMedPDF
• 7. Chung SM, Chang MC. Assessment of the information provided by ChatGPT regarding exercise for patients with type 2 diabetes: a pilot study. BMJ Health Care Inform 2024;31:e101006.ArticlePubMedPMC
• 8. Yang S, Chang MC. The assessment of the validity, safety, and utility of ChatGPT for patients with herniated lumbar disc: a preliminary study. Medicine (Baltimore) 2024;103:e38445.ArticlePubMedPMC

Figure & Data

References

Citations

Citations to this article as recorded by