Postural analysis in ventral and dorsal decubitus babies using deep learning techniques: a protocol study

Abstract

The analysis of posture in the early stages of motor development has always been a subject of research and study. With the evolution of new technologies, the need arises to implement evaluation tools that allow an objective and effective assessment of postural control, which is intrinsically linked to motor development. Objectives: The objective was to analyze posture in babies from 0 to 6 months in ventral and dorsal decubitus using artificial intelligence to determine objective parameters of postural assessment. Methods: The study is an observational and cross-sectional study. The babies will be studied following a systematic kinesiological assessment, and the images of the babies will be taken, both in ventral and dorsal decubitus, on a glass platform, to analyze their posture by means of deep learning techniques. Results: Many authors have investigated posture in newborns. However, there is no method for assessing motor and postural development to determine the support area of typically developing babies. Artificial intelligence is postulated as an effective tool to objectively analyze the posture of babies and detect possible delays. Using deep learning techniques as a predictive tool, the support areas of each baby will be defined according to their age. Conclusions: Early detection of motor or postural developmental delays in babies to optimize effective treatment is of great importance. Artificial intelligence can help manage the complexity and growing volume of data in healthcare by knowing the correct postural control at each stage of a baby’s early months, while reducing the workload of healthcare professionals by facilitating decision-making.