Foundations of Multiparametric Brain Tumour Imaging Characterisation Using Machine Learning

Jian, Anne; Jang, Kevin; Russo, Carlo; Liu, Sidong; Ieva, Antonio Di

doi:10.1007/978-3-030-85292-4_22

Cited by 12 publications

(2 citation statements)

References 53 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…BNN is a mathematical calculation technique with a formula that determines each layer. BNN is a type of artificial neural network (ANN) using the supervised learning method (supervised learning) [39], [40]. In supervised learning, there is a pair of input and output data that is used to train the artificial neural network until the desired weight is obtained.…”

Section: G Backpropagation Neural Network (Bnn)mentioning

confidence: 99%

Analysis and Improvement of CT-Scan/MRI Medical Image Quality in Stroke Patients with Hybrid Thresholding Method: A Case Study in Padang BMC Hospital

Sumijan¹,

Purnama²

2023

Int. J. Adv. Sci. Eng. Inf. Technol.

View full text Add to dashboard Cite

Stroke is one of the main problems in the medical field, one of the top 3 causes of death in the world. 1) Coronary heart as much as 13%, 2) Cancer as much as 12%, 3) Stroke as much as 7.9%. Stroke often occurs in developing countries, such as Indonesia. One effort to minimize the risk of stroke is to take preventive measures in stroke patients, both before and after a stroke. The way to prevent stroke is by primary and secondary measures. Prevention of stroke symptoms can be avoided with fast and appropriate treatment in stroke care according to medical service standards. This study applies texture analysis and improves the quality of CT-Scan or Magnetic Resonance Imaging (MRI) brain images using the Hybrid Thresholding (HT) method, which is a combination of edge detection methods and P-Files. Differences in the texture of images of brain hemorrhages that indicate stroke and clarify the quality of images of brain hemorrhages that indicate strokes with the parameters of contrast, correlation, energy, and homogeneity. The results of segmentation and feature extraction and texture are then classified using the Backprogation Neural Network (BNN) method with variations in learning rate values. This study produces the best test at a learning rate of 0.1 with an error percentage of 2%. The results of the classification calculation of the area of brain hemorrhage and analysis of image quality in classifying stroke brain with an accuracy level of 27/3 x 100% = 90%.

show abstract

Section: G Backpropagation Neural Network (Bnn)mentioning

confidence: 99%

Analysis and Improvement of CT-Scan/MRI Medical Image Quality in Stroke Patients with Hybrid Thresholding Method: A Case Study in Padang BMC Hospital

Sumijan¹,

Purnama²

2023

Int. J. Adv. Sci. Eng. Inf. Technol.

View full text Add to dashboard Cite

show abstract

“…Based on the size and nature of data set, and the learning problem at hand, various statistical, ML, and DL models have been selected with their own advantages and disadvantages (Table 2). 94 Their performance has been compared with different outcomes 12,14,18,32,40,45,50,57,67,70 ; however, simple statistical models were superior in some cases. 39,52,54 Molina-Garcia et al 39 showed that a linear model built on a small set of morphological variables performed equally well to ML methods.…”

Section: Models and Evaluationmentioning

confidence: 99%

Artificial Intelligence for Survival Prediction in Brain Tumors on Neuroimaging

2022

Self Cite

View full text Add to dashboard Cite

Survival prediction of patients affected by brain tumors provides essential information to guide surgical planning, adjuvant treatment selection, and patient counseling. Current reliance on clinical factors, such as Karnofsky Performance Status Scale, and simplistic radiological characteristics are, however, inadequate for survival prediction in tumors such as glioma that demonstrate molecular and clinical heterogeneity with variable survival outcomes. Advances in the domain of artificial intelligence have afforded powerful tools to capture a large number of hidden high-dimensional imaging features that reflect abundant information about tumor structure and physiology. Here, we provide an overview of current literature that apply computational analysis tools such as radiomics and machine learning methods to the pipeline of image preprocessing, tumor segmentation, feature extraction, and construction of classifiers to establish survival prediction models based on neuroimaging. We also discuss challenges relating to the development and evaluation of such models and explore ethical issues surrounding the future use of machine learning predictions.

show abstract