A novel classification method for adult spinal deformity diseases using deep learning of gait data

Tsukuba, Japan—Patients with adult spinal deformity (ASD) have altered gait patterns because of the spinal deformity; hence, gait analysis may be effective for diagnosis. However, the conventional methods of gait analysis may be inadequate for studying the characteristicsof posture and movement during walking, which are essential for diagnosis. Recently, deep learning technology using video images has been used.

Using this technique, researchers have developed a new method to accurately capture the rhythm and symmetry of body movements during walking, which may be used to classify the periodicity and postures adopted by the lower extremities and the body during gait. We tested this method using walking videos of 81 patients and achieved a correct response rate of 71.43%, which was more accurate than the conventional method (66.30%), and confirmed its effectiveness for diagnosing ASD.

In the future, this technique may allow real-time analysis of moving images in clinical settings to enable instantaneous confirmation and rapid diagnosis of ASD.

###
This work was partly supported by AMED under Grant Number JP23YM0126803.

_{Original Paper}

Title of original paper:
PhaseMix: A Periodic Motion Fusion Method for Adult Spinal Deformity Classification

Journal:
IEEE Access

DOI:
10.1109/ACCESS.2024.3479165

_{Correspondence}

Professor KURODA, Yoshihiro
Institute of Systems and Information Engineering, University of Tsukuba

Kaixu Chen
Degree Programs in Systems and Information Engineering, Graduate School of Science and Technology, University of Tsukuba

Assistant Professor MIURA, Kousei
Department of Orthopaedic Surgery, Institute of Medicine, University of Tsukuba

_{Related Link}

Institute of Systems and Information Engineering