Validation of hand and foot anatomical feature measurements from smartphone images

摘要

A smartphone mobile medical application, previously presented as a tool for individuals with hand arthritis to assess and monitor the progress of their disease, has been modified and expanded to include extraction of anatomical features from the hand (joint/finger width, and angulation) and foot (length, width, big toe angle, and arch height index) from smartphone camera images. Image processing algorithms and automated measurements were validated by performing tests on digital hand models, rigid plastic hand models, and real human hands and feet to determine accuracy and reproducibility compared to conventional measurement tools such as calipers, rulers, and goniometers. The mobile application was able to provide finger joint width measurements with accuracy better than 0.34 (±0.25) millimeters. Joint angulation measurement accuracy was better than 0.50 (±0.45) degrees. The automatically calculated foot length accuracy was 1.20 (±1.27) millimeters and the foot width accuracy was 1.93 (±1.92) millimeters. Hallux valgus angle (used in assessing bunions) accuracy was 1.30 (±1.29) degrees. Arch height index (AHI) measurements had an accuracy of 0.02 (±0.01). Combined with in-app documentation of symptoms, treatment, and lifestyle factors, the anatomical feature measurements can be used by both healthcare professionals and manufacturers. Applications include: diagnosing hand osteoarthritis; providing custom finger splint measurements; providing compression glove measurements for burn and lymphedema patients; determining foot dimensions for custom shoe sizing, insoles, orthotics, or foot splints; and assessing arch height index and bunion treatment effectiveness.