RESUMO
Direct observation is a ground-truth measure for physical behavior, but the high cost limits widespread use. The purpose of this study was to develop and test machine learning methods to recognize aspects of physical behavior and location from videos of human movement: Adults (N = 26, aged 18-59 y) were recorded in their natural environment for two, 2- to 3-h sessions. Trained research assistants annotated videos using commercially available software including the following taxonomies: (1) sedentary versus non-sedentary (two classes); (2) activity type (four classes: sedentary, walking, running, and mixed movement); and (3) activity intensity (four classes: sedentary, light, moderate, and vigorous). Four machine learning approaches were trained and evaluated for each taxonomy. Models were trained on 80% of the videos, validated on 10%, and final accuracy is reported on the remaining 10% of the videos not used in training. Overall accuracy was as follows: 87.4% for Taxonomy 1, 63.1% for Taxonomy 2, and 68.6% for Taxonomy 3. This study shows it is possible to use computer vision to annotate aspects of physical behavior, speeding up the time and reducing labor required for direct observation. Future research should test these machine learning models on larger, independent datasets and take advantage of analysis of video fragments, rather than individual still images.
Assuntos
Computadores , Trabalho de Parto , Adulto , Humanos , Gravidez , Feminino , Software , Meio Ambiente , Aprendizado de MáquinaRESUMO
With a growing aging population, the routine assessment of physical function may become a critical component of clinical practice. The purpose of this cross-sectional study is to compare two common assessments of muscular function: (1) isometric knee extension strength (KES) and (2) sit-to-stand (STS) muscle power tests, in predicting objective physical function (i.e., gait speed) in aging adults. 84 adults (56% female, mean (SD) age = 66.6 (9.4) years) had their relative KES, STS power, usual gait speed (UGS), and fast gait speed (FGS) assessed. Multiple linear regression examined the associations between KES, STS power, and gait outcomes. When entered in separate models, KES and STS power were both independently associated with UGS and FGS (Std. ß = 0.35-0.44 and 0.42-0.55 for KES and STS power, respectively). When entered in the same model, STS power was associated with UGS and FGS (Std. ß = 0.37 [95%CI: 0.15, 0.58] and 0.51 [95%CI: 0.31, 0.70], respectively), while KES was only associated with FGS (Std. ß = 0.25 [95%CI: 0.02, 0.48]). STS power seems to be a valid indicator of function in aging adults. Its feasibility as a screening tool for "low" function in the primary care setting should be explored.