Semi-supervised lung nodule retrieval

원문 초록

Content based image retrieval (CBIR) provides the clinician with visual information that can support, and hopefully improve, his or her decision making process. Given an input query image, a CBIR system provides as its output a set of images, ranked by similarity to the query image. Retrieved images may come with relevant information, such as biopsy-based malignancy labeling, or categorization. Ground truth on similarity between dataset elements (e.g. between nodules) is not readily available, thus greatly challenging machine learning methods. Such annotations are particularly difficult to obtain, due to the subjective nature of the task, with high inter-observer variability requiring multiple expert annotators. Consequently, past approaches have focused on manual feature extraction, while current approaches use auxiliary tasks, such as a binary classification task (e.g. malignancy), for which ground-true is more readily accessible. However, in a previous study, we have shown that binary auxiliary tasks are inferior to the usage of a rough similarity estimate that are derived from data annotations. The current study suggests a semi-supervised approach that involves two steps: 1) Automatic annotation of a given partially labeled dataset; 2) Learning a semantic similarity metric space based on the predicated annotations. The proposed system is demonstrated in lung nodule retrieval using the LIDC dataset, and shows that it is feasible to learn embedding from predicted ratings. The semi-supervised approach has demonstrated a significantly higher discriminative ability than the fully-unsupervised reference.