Research On Information Bottleneck Attribution Based Retinal Disease Classification Using OCT Image

The existence of deep learning’s "black box" makes it difficult to understand how the algorithms analyze patterns and make image-level predictions.A representation of the pixels contributing the most to the algorithm’s classification requires new insights.Classification methods for neurodegenerative ocular disorders have been developed with considerable efficacy using machine learning and image processing techniques.However,the techniques’ robustness and transferability remain uncertain.A new classification method has been developed based upon the information bottleneck to analyze the attribution of each feature and the information provided for the network prediction in each input area for a clear understanding of the affectability of a blackbox model.This study aims to identify the most significant facts to create an accurate forecast based on the above research findings.Attribution scores for characteristics are generated based on different algorithms,explaining how specialized prediction is carried out at the instance level.To maximize the amount of information and ensure that zero-bit areas are not crucial for predicting,an Information Bottleneck(IB)has been implemented as an attribution principle.In general,deep learning-based automation methods have been developed and successfully addressed the main weaknesses of traditional systems based on manual feature extraction.In this research,the attribution information bottleneck is applied to limit the information flow and assess the amount of information image areas produced in bits by allowing noise to intermediate feature maps.Our studies indicate that the information bottleneck for attribution(IBA)has increased the model interpretability and gives a more reliable estimation on a publicly available dataset.