Given the scarcity of sufficient annotated data, using small sets of labeled samples under semi-supervision in biomedical imaging becomes necessary. Despite being highly successful, deep learning algorithms demand plenty of data to obtain significant performance. Complex data models make the usage of these methods costly. Selecting the correct model and tuning the hyperparameters of a model are also difficult jobs. Hence, a novel approach namely affinity propagation-based semi-supervised segmentation (APSS) is proposed. Here, affinity propagation clustering is modified and integrated with the advanced learning techniques that can efficiently use limited training data by discarding the completely exploited labeled data points. Moreover, a novel affinity calculation method is proposed considering both the Euclidean and geodesic distances to compute the distance between the two points on the histogram. This twofold contribution is tested using the three standard datasets (the International Skin Imaging Collaboration (ISIC) dermoscopic image dataset, the retinal fundus image dataset, and the liver tumor segmentation (LiTS) dataset). Results are compared with the three standard semi-supervised algorithms and four supervised algorithms. The effectiveness of the APSS approach in finding and exploiting the relationship between the labeled and unlabeled datasets is demonstrated in terms of qualitative (subjective evaluation and visual inspection) and quantitative performance (objective evaluation and numerical measurements).

A global pandemic scenario is witnessed worldwide owing to the menace of the rapid outbreak of the deadly COVID-19 virus. To save mankind from this apocalyptic onslaught, it is essential to curb the fast spreading of this dreadful virus. Moreover, the absence of specialized drugs has made the scenario even more badly and thus an early-stage adoption of necessary precautionary measures would provide requisite supportive treatment for its prevention. The prime objective of this article is to use radiological images as a tool to help in early diagnosis. The interval type 2 fuzzy clustering is blended with the concept of superpixels, and metaheuristics to efficiently segment the radiological images. Despite noise sensitivity of watershed-based approach, it is adopted for superpixel computation owing to its simplicity where the noise problem is handled by the important edge information of the gradient image is preserved with the help of morphological opening and closing based reconstruction operations. The traditional objective function of the fuzzy c-means clustering algorithm is modified to incorporate the spatial information from the neighboring superpixel-based local window. The computational overhead associated with the processing of a huge amount of spatial information is reduced by incorporating the concept of superpixels and the optimal clusters are determined by a modified version of the flower pollination algorithm. Although the proposed approach performs well but should not be considered as an alternative to gold standard detection tests of COVID-19. Experimental results are found to be promising enough to deploy this approach for real-life applications.