ABSTRACT

Our main objective is to identify the early stages of lung disease, but also to compare this efficacy using several machine teaching methods. After a very thorough examination of the material, the researchers discovered that some classifications have poor performance, while others have better accuracy, but usually difficult to obtain close to 100%. Due to improper manipulation of DICOM images, there is usually a reduced level of accuracy, but also a relatively considerable advantage of complexity. Many various forms such as pictures are used in interpreting health care images, however, scanners are usually chosen as they have reduced clutter. With regard to health care imaging analysis, lung nodule identification, but also classifications, highlight collection, in particular staging predictions of intestinal malignancy, a deeper understanding has been shown that provides the most effective technique overall. For distinct bronchial areas throughout this same initial step with this methodology, information manipulation technologies have been applied. K Means are used to classify information data. Relevant characteristics were also taken from segmentation images, while identification is performed using a wide variety of machine intelligence algorithms. Efficiency, sensitivities, in particular, as well as identification time, were used to assess the effectiveness of proposed techniques. Keywords: Machine Learning; Machine Learning; DICOM images; Segmentation
Publication date: 01/11/2021

https://ijbpas.com/pdf/2021/November/MS_IJBPAS_2021_NOV_SPCL1092.pdf

Download PDF
https://doi.org/10.31032/IJBPAS/2021/10.11.1092