Artificial Intelligence–Based Mobile Application for Identifying Suitable Height Range of High Heels

Abstract

The wearing of overly high high-heeled shoes can cause irreversible physiological damage. Doctors typically analyze the calcaneus when an individual is standing on tiptoes to observe changes in the Achilles tendon to determine the suitable high-heel height range. A mobile application system for evaluating the maximum heel height suitable for women was developed in this study. The system input is a smartphone-recorded video of an individual’s while standing on tiptoes. The trained artificial intelligence (AI) evaluation system outputs the maximum heel height suitable for the wearer. This study used three core technologies to implement this AI evaluation system. First, EfficientDet was used to realize object detection, and area of the left heel, right heel, and a red cube were bounded. These bounding region pictures were then input into the convolutional neural network (CNN) model for binary classification; the accuracy rate of the integrated EfficientDet and CNN models was 0.86. Finally, the mode algorithm was applied to correct errors in model evaluation and improve overall accuracy. The system developed in this study can evaluate the maximum heel height of a high-heeled shoe with a mean absolute error of 0.8 cm. Only the operation of a smartphone is required without additional platform assistance. Thus, this system has high practicability and usability.

Keywords

• Height of high heel
• artificial intelligence (AI)
• EfficientDet
• convolutional neural network (CNN)
• mobile application