Pelican optimized extreme learning machine based prognosis of heart failure using textural patterns in CMR images

One of the abnormal cardiovascular conditions with the greatest rate of increase is heart failure (HF). There is a lot of correlation in the exterior symptoms observed that are typically ignored, particularly when there is progression from one stage to another. Planning an efficient treatment requires grouping patients into different phenotypic categories. In CMR imaging, these conditions show damaged heart muscles. Understanding the classification performance of ELM aided by optimal texture features is the main aim of the proposed work. In this study, Krawtchouk moment and the co-occurrence of neighboring sparse local ternary pattern descriptors are used to extract texture and anatomical information of LV. However, an effective feature selection strategy is required to handle the high dimensionality of the feature vector. The primary benefit of the Pelican optimization algorithm (POA), which takes inspiration from nature, is its capacity to carry out both global and local searches effectively. This paper suggests usage of multi-objective POA to optimize the proportion of extracted features and improve the performance of the ELM classifier. The people with various HF stages have been diagnosed with the maximum accuracy of 95.7 % as a result of this inclusion. The enhanced traits have distinguished moderate and severe HF patients from control participants with significant performance. The proposed work further shows the impact of tissue characterization from CMR image and optimized features-based classification on the HF stage detection.