Publications by authors named "Teresa M Correia"
In this work, we propose a model-based deep learning reconstruction algorithm for optical projection tomography (ToMoDL), to greatly reduce acquisition and reconstruction times. The proposed method iterates over a data consistency step and an image domain artefact removal step achieved by a convolutional neural network. A preprocessing stage is also included to avoid potential misalignments between the sample center of rotation and the detector.
View Article and Find Full Text PDFArticle Synopsis
- A key method for evaluating heart health after a heart attack (myocardial infarction) is using delayed enhancement MRI (DE-MRI), which identifies viable versus nonviable heart tissue effectively.
- The EMIDEC challenge aimed to assess whether deep learning techniques could differentiate between healthy and pathological heart exams, and also to automatically calculate the extent of myocardial injury using a dataset of 150 MRI cases.
- Results showed that high accuracy (up to 0.92) in classifying heart exams is achievable, but improvements are needed for accurately segmenting the affected areas due to their small size and low contrast with surrounding tissues.