INTEGRATING SEMI-SUPERVISED DISCRIMINATIVE CLASSIFICATION WITH DENOISING FOR ENHANCED STATISTICAL ANALYSIS IN MEDICAL IMAGE PROCESSING
INTEGRATING SEMI-SUPERVISED DISCRIMINATIVE CLASSIFICATION WITH DENOISING FOR ENHANCED STATISTICAL ANALYSIS IN MEDICAL IMAGE PROCESSING
Where did the research take place?
The study site has not been established. Author addresses may differ from where the research occurred.
A plain-language reading has not been prepared for this paper yet.
Original abstract
Despite the superior advantages of discriminative approaches in generalizing, there are challenges faced in real-world applications such as medical image analysis. Existence of robustness to outliers and noise in predictor values is a test in which one traditional incomplete attention is given in existing literature. We state that the denoising process can be refined to a considerable extent with the help of training models on both the labeled and unlabeled data; in this way, it is possible to obtain a more accurate representation of the natural structure of the sample manifold. The proposed method, the suggested paper is a new semi-supervised robust discriminative classification with its foundations in the least-squares version of the linear discriminant analysis. Its overall objective is the joint discovery of features-noises and sample- outliers via the joint use of labeled training data and unlabeled testing data. Sufficient experimentation was performed based on synthetic data, semi-supervised learning benchmarks as well as two datasets specifically centered around the diagnosis of neurodegenerative diseases (Parkinson and Alzheimer). The medical imaging data is characterized by inherent noise, so the advancement of efficient machine learning tools is the key to proper diagnosis of neurodegenerative diseases. These findings demonstrate that the given method is invariably an improvement to the baseline models and other state-of-the-art approaches in terms of the accuracy and the area under the ROC curve. Our study bridges the research gap of statistics with a solid discriminative approach that aims at addressing the outlier and noise present in image analysis in medicine. This attempt has the significant promise of improving neurodegenerative disease diagnosis.