Convolutional neural networks (CNNs) have been popularly used to solve the problem of cell/nuclei classi cationand segmentation in histopathology images. Despite their pervasiveness, CNNs are ne-tuned on speci c, largeand labeled datasets as these datasets are hard to collect and annotate. However, this is not a scalable approach.In this work, we aim to gain deeper insights into the nature of the problem. We used a cervical cancer datasetwith cells labeled into four classes by an expert pathologist. By employing pre-training on this dataset, wepropose a one-shot learning model for cervical cell classification in histopathology tissue images. We extractregional maximum activation of convolutions (R-MAC) global descriptors and train a one-shot learning memorymodule with the goal of using it for various cancer types and eliminate the need for expensive, diffcult to collect,large, labeled whole slide image (WSI) datasets. Our model achieved 94.6% accuracy in detecting the four cellclasses on the test dataset. Further, we present our analysis of the dataset and features to better understandand visualize the problem in general.
展开▼
