No description available.

No description available.

Breast cancer is one of the most common cancers in women, affecting hundreds of women. Even though the detection of cancer has been largely studied, the decision of which strategy to take concerning oncoplastic surgery still relies almost exclusively on the surgeon's perception of post-surgical aesthetic result, which sometime leads to unsatisfactory outcomes. In order to empower the patients on the joint decision process there needs to exist a better communication between the parts. This can be achieved by developing medical grade 3D models of the breast and explaining better the surgical options and their results. In order to obtain such models, some effort has been made concerning multi-modality radiological imaging combination. This line of research has yet to mature. In turn, the modality alignment requires accurate landmarks to be produced. 2D Ultrasound imaging has not been sufficiently studied for multimodal registration due to the image characteristics and thus, landmark segmentation is of utmost importance. This task can be challenging since US data presents high specular noise levels and the presence of some tissues alters the perception of other tissues. Objectives: ● Study and evaluation of different techniques for anatomical landmark segmentation, such as Skin, Fat and Glandular tissue, Lesions (masses and cysts), Pectoral muscle; ● Development of Ultrasound segmentation methods for acquiring landmarks; ● Evaluation of the developed methods with manual annotations and comparison of results with the current algorithm alternatives. / Breast cancer is one of the most common cancers in women, affecting hundreds of women. Even though the detection of cancer has been largely studied, the decision of which strategy to take concerning oncoplastic surgery still relies almost exclusively on the surgeon's perception of post-surgical aesthetic result, which sometime leads to unsatisfactory outcomes. In order to empower the patients on the joint decision process there needs to exist a better communication between the parts. This can be achieved by developing medical grade 3D models of the breast and explaining better the surgical options and their results. In order to obtain such models, some effort has been made concerning multi-modality radiological imaging combination. This line of research has yet to mature. In turn, the modality alignment requires accurate landmarks to be produced. 2D Ultrasound imaging has not been sufficiently studied for multimodal registration due to the image characteristics and thus, landmark segmentation is of utmost importance. This task can be challenging since US data presents high specular noise levels and the presence of some tissues alters the perception of other tissues. Objectives: ● Study and evaluation of different techniques for anatomical landmark segmentation, such as Skin, Fat and Glandular tissue, Lesions (masses and cysts), Pectoral muscle; ● Development of Ultrasound segmentation methods for acquiring landmarks; ● Evaluation of the developed methods with manual annotations and comparison of results with the current algorithm alternatives.

No description available.

Deep Learning models, in particular Convolutional Neural Networks, have become the state-of-the-art in different domains, such as image classification, object detection and other computer vision tasks. However, despite their overwhelming predictive performance, they are still, for the most part, considered black-boxes, making it difficult to understand the reasoning behind their outputted decisions. As such, and with the growing interest in deploying such models into real world scenarios, the need for explainable systems has arisen. Therefore, this dissertation tries to mitigate this growing need, by proposing a novel CNN architecture, composed of an explainer and a classifier. The network, trained end-to-end, constitutes an in-model explainability method, that not only outputs decisions as well as visual explanations of what the network is focusing on to produce such decisions.

No description available.

No description available.

No description available.

No description available.

No description available.