João Manuel Pedrosa

Abstract
Chest X-ray scans are frequently requested to detect the presence of abnormalities, due to their low-cost and non-invasive nature. The interpretation of these images can be automated to prioritize more urgent exams through deep learning models, but the presence of image artifacts, e.g. lettering, often generates a harmful bias in the classifiers and an increase of false positive results. Consequently, healthcare would benefit from a system that selects the thoracic region of interest prior to deciding whether an image is possibly pathologic. The current work tackles this binary classification exercise, in which an image is either normal or abnormal, using an attention-driven and spatially unsupervised Spatial Transformer Network (STERN), that takes advantage of a novel domain-specific loss to better frame the region of interest. Unlike the state of the art, in which this type of networks is usually employed for image alignment, this work proposes a spatial transformer module that is used specifically for attention, as an alternative to the standard object detection models that typically precede the classifier to crop out the region of interest. In sum, the proposed end-to-end architecture dynamically scales and aligns the input images to maximize the classifier's performance, by selecting the thorax with translation and non-isotropic scaling transformations, and thus eliminating artifacts. Additionally, this paper provides an extensive and objective analysis of the selected regions of interest, by proposing a set of mathematical evaluation metrics. The results indicate that the STERN achieves similar results to using YOLO-cropped images, with reduced computational cost and without the need for localization labels. More specifically, the system is able to distinguish abnormal frontal images from the CheXpert dataset, with a mean AUC of 85.67% -a 2.55% improvement vs. the 0.98% improvement achieved by the YOLO-based counterpart in comparison to a standard baseline classifier. At the same time, the STERN approach requires less than 2/3 of the training parameters, while increasing the inference time per batch in less than 2 ms. Code available via GitHub.

Abstract
Chest radiography has been widely used for automatic analysis through deep learning (DL) techniques. However, in the manual analysis of these scans, comparison with images at previous time points is commonly done, in order to establish a longitudinal reference. The usage of longitudinal information in automatic analysis is not a common practice, but it might provide relevant information for desired output. In this work, the application of longitudinal information for the detection of cardiomegaly and change in pairs of CXR images was studied. Multiple experiments were performed, where the inclusion of longitudinal information was done at the features level and at the input level. The impact of the alignment of the image pairs (through a developed method) was also studied. The usage of aligned images was revealed to improve the final mcs for both the detection of pathology and change, in comparison to a standard multi-label classifier baseline. The model that uses concatenated image features outperformed the remaining, with an Area Under the Receiver Operating Characteristics Curve (AUC) of 0.858 for change detection, and presenting an AUC of 0.897 for the detection of pathology, showing that pathology features can be used to predict more efficiently the comparison between images. In order to further improve the developed methods, data augmentation techniques were studied. These proved that increasing the representation of minority classes leads to higher noise in the dataset. It also showed that neglecting the temporal order of the images can be an advantageous augmentation technique in longitudinal change studies.

Abstract
Echocardiography is the most frequently used imaging modality in cardiology. However, its acquisition is affected by inter-observer variability and largely dependent on the operator's experience. In this context, artificial intelligence techniques could reduce these variabilities and provide a user independent system. In recent years, machine learning (ML) algorithms have been used in echocardiography to automate echocardiographic acquisition. This review focuses on the state-of-the-art studies that use ML to automate tasks regarding the acquisition of echocardiograms, including quality assessment (QA), recognition of cardiac views and assisted probe guidance during the scanning process. The results indicate that performance of automated acquisition was overall good, but most studies lack variability in their datasets. From our comprehensive review, we believe automated acquisition has the potential not only to improve accuracy of diagnosis, but also help novice operators build expertise and facilitate point of care healthcare in medically underserved areas.

Abstract
Chest radiography is one of the most common imaging exams, but its interpretation is often challenging and timeconsuming, which has motivated the development of automated tools for pathology/abnormality detection. Deep learning models trained on large-scale chest X-ray datasets have shown promising results but are highly dependent on the quality of the data. However, these datasets often contain incorrect metadata and non-compliant or corrupted images. These inconsistencies are ultimately incorporated in the training process, impairing the validity of the results. In this study, a novel approach to detect non-compliant images based on deep features extracted from a patient position classification model and a pre-trained VGG16 model are proposed. This method is applied to CheXpert, a widely used public dataset. From a pool of 100 images, it is shown that the deep feature-based methods based on a patient position classification model are able to retrieve a larger number of non-compliant images (up to 81% of non-compliant images), when compared to the same methods but based on a pretrained VGG16 (up to 73%) and the state of the art uncertainty-based method (50%).

Abstract
The importance of X-Ray imaging analysis is paramount for healthcare institutions since it is the main imaging modality for patient diagnosis, and deep learning can be used to aid clinicians in image diagnosis or structure segmentation. In recent years, several articles demonstrate the capability that deep learning models have in classifying and segmenting chest x-ray images if trained in an annotated dataset. Unfortunately, for segmentation tasks, only a few relatively small datasets have annotations, which poses a problem for the training of robust deep learning strategies. In this work, a semi-supervised approach is developed which consists of using available information regarding other anatomical structures to guide the segmentation when the groundtruth segmentation for a given structure is not available. This semi-supervised is compared with a fully-supervised approach for the tasks of lung segmentation and for multi-structure segmentation (lungs, heart and clavicles) in chest x-ray images. The semi-supervised lung predictions are evaluated visually and show relevant improvements, therefore this approach could be used to improve performance in external datasets with missing groundtruth. The multi-structure predictions show an improvement in mean absolute and Hausdorff distances when compared to a fully supervised approach and visual analysis of the segmentations shows that false positive predictions are removed. In conclusion, the developed method results in a new strategy that can help solve the problem of missing annotations and increase the quality of predictions in new datasets.