Abstract

Deep convolutional networks have proved to be extremely powerful in image analysis. However.they tend to be biased toward local features such as texture and often fail to capture the globalstructure of image and volume data. Persistent homology (PH), a tool from the emerging field oftopological data analysis, has been successfully used to detect global characteristics of data thatare overlooked by conventional methods. We present Cubical Ripser, an open-source softwaredesigned for high-efficiency computation of PH of cubical complexes. We will discuss how PH canbe integrated into a standard image analysis pipeline with some practical applications.