Advances in Automated Size Estimation and Geometric Feature Computation in Medical Imaging Using Machine Learning: A Comprehensive Survey

Resumen

Recently, the integration of machine learning (ML) and deep learning (DL) with
medical imaging has improved the accuracy, efficiency and automatization of disease
detection and diagnosis.Recent investigations have shown that hybrid models, ensem
ble learning, and multi-modal data fusion are effective diagnostic for vast range of
diagnosis and disorder detection. The traditional ML and deep learning model com
prises of feature selection, hyperparameter optimization, and automated machine learn
ing (AutoML), exhibiting significant and reliable outcomes. The advancement with
transformer-based models, explainable AI, and automated end-to-end exhibits promis
ing clinical translation and personalized healthcare. Despite those developments, lack
of standardization of metrics of evaluation, cross-domain generalization, and the im
plementation of AI solutions into clinical practice. This review paper provides a com
prehensive analysis of the existing approaches for future development of research, and
highlights the opportunities of AI-based automation to revolutionize medical imaging.
Additionally, this paper aimed to improve its precision, accuracy, and detection, as well
as classification and quantification of different diseases. The reviewed methodically an
alyzes the state-of-the-art methods, based on geometric feature extraction, radiomics,
size computation with AutoML, and anomaly detection in medical images. The anal
ysis is based on the hybrid model, ensemble learning, and multi-modal data fusion in
improving the diagnostic accuracy with particular focus on disorder analysis. Compar
ative research identifies the advantages limitations of various methods, and the new
trends of explainable AI, transformer-based models, and automated model indicate the opportunities of clinical translation. The present paper presented detailed discussion of modern methods, emphasizing geometry feature calculation, radiomics, automated
size determination, and anomaly detection in various imaging modalities, such as MRI,
CT, ultrasound, and histopathological imaging.