Feature Extraction and Selection Applied to Bone Radiographs in Traumatological Surgical Procedures: A Quasi-Survey

The current medical field faces the challenge of interpreting large amounts of clinical data, including radiographic images, in a timely and accurate manner. This article seeks to carry out a bibliographical survey involving the combined use of Feature Extraction (FE) and Feature Selection (FS) techniques in bone radiographs by proposing stages of an artificial intelligence model and suggesting materials for surgical procedures. The feature extraction technique involves identifying and extracting relevant features from large datasets and converting them into a more meaningful and compact form for modelling and analysis. Feature selection is employed to choose a subset of pertinent features from a more extensive set that aims to enhance machine learning models’ performance by reducing the total number of features and eliminating those deemed irrelevant or redundant. Using these techniques to automate the interpretation of radiographic images can save costs, time, and resources, especially in large sets of images data, and contribute to better learning and classification of the proposed model.