Specular reflections (i.e., highlight) always exist in endoscopic images, and they can severely disturb surgeons' observation and judgment. In an augmented reality (AR)-based surgery navigation system, the highlight may also lead to the failure of feature extraction or registration. In this paper, we propose an adaptive robust principal component analysis (Adaptive-RPCA) method to remove the specular reflections in endoscopic image sequences. It can iteratively optimize the sparse part parameter during RPCA decomposition. In this new approach, we first adaptively detect the highlight image based on pixels. With the proposed distance metric algorithm, it then automatically measures the similarity distance between the sparse result image and the detected highlight image. Finally, the low-rank and sparse results are obtained by enforcing the similarity distance between the two types of images to fall within a certain range. Our method has been verified by multiple different types of endoscopic image sequences in minimally invasive surgery (MIS). The experiments and clinical blind tests demonstrate that the new Adaptive-RPCA method can obtain the optimal sparse decomposition parameters directly and can generate robust highlight removal results. Compared with the state-of-the-art approaches, the proposed method not only achieves the better highlight removal results but also can adaptively process image sequences.

MOTIVATION: RNA-seq technology has been widely adopted as an attractive alternative to microarray-based methods to study global gene expression. However, robust statistical tools to analyze these complex datasets are still lacking. By grouping genes with similar expression profiles across treatments, cluster analysis provides insight into gene functions and networks, and hence is an important technique for RNA-seq data analysis. RESULTS: In this manuscript, we derive clustering algorithms based on appropriate probability models for RNA-seq data. An expectation-maximization algorithm and another two stochastic versions of expectation-maximization algorithms are described. In addition, a strategy for initialization based on likelihood is proposed to improve the clustering algorithms. Moreover, we present a model-based hybrid-hierarchical clustering method to generate a tree structure that allows visualization of relationships among clusters as well as flexibility of choosing the number of clusters. Results from both simulation studies and analysis of a maize RNA-seq dataset show that our proposed methods provide better clustering results than alternative methods such as the K-means algorithm and hierarchical clustering methods that are not based on probability models. AVAILABILITY AND IMPLEMENTATION: An R package, MBCluster.Seq, has been developed to implement our proposed algorithms. This R package provides fast computation and is publicly available at http://www.r-project.org