Toward automating the diagnosis of gastrointestinal parasites in cats and dogs.

Diagnosing gastrointestinal parasites by microscopy slide examination often leads to human interpretation errors, which may occur due to fatigue, lack of training and infrastructure, presence of artifacts (e.g., various types of cells, algae, yeasts), and other reasons. We have investigated the stages in automating the process to cope with the interpretation errors. This work presents advances in two stages focused on gastrointestinal parasites of cats and dogs: a new parasitological processing technique, named TF-Test VetPet, and a microscopy image analysis pipeline based on deep learning methods. TF-Test VetPet improves image quality by reducing cluttering (i.e., eliminating artifacts), which favors automated image analysis. The proposed pipeline can identify three species of parasites in cats and five in dogs, distinguishing them from fecal impurities with an average accuracy of 98,6%. We also make available the two datasets with images of parasites of dogs and cats, which were obtained by processing fecal smears with temporary staining using TF-Test VetPet.

An Approach for Asbestos-related Pleural Plaque Detection.

Asbestos is a toxic ore widely used in construction and commercial products. Asbestos tends to dissolve into fibers and after years inhaling them, these fibers calcify and form plaques on the pleura. Despite being benign, pleural plaques may indicate an immunologic deficiency or dysfunctional lung areas. We propose a pipeline for asbestos-related pleural plaque detection in CT images of the human thorax based on the following operations: lung segmentation, 3D patch selection along the pleura, a convolutional neural network (CNN) for feature extraction, and classification by support vector machines (SVM). Due to the scarcity of publicly available and annotated datasets of pleural plaques, the proposed CNN relies on architecture learning with random weights obtained by a PCA-based approach instead of using traditional filter learning by backpropagation. Experiments show that the proposed CNN can outperform its counterparts based on backpropagation for small training sets.