The increasing popularity of attention mechanisms in deep learning algorithms for computer vision and natural language processing made these models attractive to other research domains. In healthcare, there is a strong need for tools that may improve the routines of the clinicians and the patients. Naturally, the use of attention-based algorithms for medical applications occurred smoothly. However, being healthcare a domain that depends on high-stake decisions, the scientific community must ponder if these high-performing algorithms fit the needs of medical applications. With this motto, this paper extensively reviews the use of attention mechanisms in machine learning methods (including Transformers) for several medical applications based on the types of tasks that may integrate several works pipelines of the medical domain. This work distinguishes itself from its predecessors by proposing a critical analysis of the claims and potentialities of attention mechanisms presented in the literature through an experimental case study on medical image classification with three different use cases. These experiments focus on the integrating process of attention mechanisms into established deep learning architectures, the analysis of their predictive power, and a visual assessment of their saliency maps generated by post-hoc explanation methods. This paper concludes with a critical analysis of the claims and potentialities presented in the literature about attention mechanisms and proposes future research lines in medical applications that may benefit from these frameworks.
Human epidermal growth factor receptor 2 (HER2) evaluation commonly requires immunohistochemistry (IHC) tests on breast cancer tissue, in addition to the standard haematoxylin and eosin (H&E) staining tests. Additional costs and time spent on further testing might be avoided if HER2 overexpression could be effectively inferred from H&E stained slides, as a preliminary indication of the IHC result. In this paper, we propose the first method that aims to achieve this goal. The proposed method is based on multiple instance learning (MIL), using a convolutional neural network (CNN) that separately processes H&E stained slide tiles and outputs an IHC label. This CNN is pretrained on IHC stained slide tiles but does not use these data during inference/testing. H&E tiles are extracted from invasive tumour areas segmented with the HASHI algorithm. The individual tile labels are then combined to obtain a single label for the whole slide. The network was trained on slides from the HER2 Scoring Contest dataset (HER2SC) and tested on two disjoint subsets of slides from the HER2SC database and the TCGA-TCIA-BRCA (BRCA) collection. The proposed method attained classification accuracy on the HER2SC test set and on the BRCA test set. Although further efforts should be devoted to achieving improved performance, the obtained results are promising, suggesting that it is possible to perform HER2 overexpression classification on H&E stained tissue slides.
Background: Breast cancer therapy improved significantly, allowing for different surgical approaches for the same disease stage, therefore offering patients different aesthetic outcomes with similar locoregional control. The purpose of the CINDERELLA trial is to evaluate an artificial-intelligence (AI) cloud-based platform (CINDERELLA platform) vs the standard approach for patient education prior to therapy. Methods: A prospective randomized international multicentre trial comparing two methods for patient education prior to therapy. After institutional ethics approval and a written informed consent, patients planned for locoregional treatment will be randomized to the intervention (CINDERELLA platform) or controls. The patients in the intervention arm will use the newly designed web-application (CINDERELLA platform, CINDERELLA APProach) to access the information related to surgery and/or radiotherapy. Using an AI system, the platform will provide the patient with a picture of her own aesthetic outcome resulting from the surgical procedure she chooses, and an objective evaluation of this aesthetic outcome (e.g., good/fair). The control group will have access to the standard approach. The primary objectives of the trial will be i) to examine the differences between the treatment arms with regards to patients’ pre-treatment expectations and the final aesthetic outcomes and ii) in the experimental arm only, the agreement of the pre-treatment AI-evaluation (output) and patient’s post-therapy self-evaluation. Discussion: The project aims to develop an easy-to-use cost-effective AI-powered tool that improves shared decision-making processes. We assume that the CINDERELLA APProach will lead to higher satisfaction, better psychosocial status, and wellbeing of breast cancer patients, and reduce the need for additional surgeries to improve aesthetic outcome.
Despite recent efforts, accuracy in group emotion recognition is still generally low. One of the reasons for these underwhelming performance levels is the scarcity of available labeled data which, like the literature approaches, is mainly focused on still images. In this work, we address this problem by adapting an inflated ResNet-50 pretrained for a similar task, activity recognition, where large labeled video datasets are available. Audio information is processed using a Bidirectional Long Short-Term Memory (Bi-LSTM) network receiving extracted features. A multimodal approach fuses audio and video information at the score level using a support vector machine classifier. Evaluation with data from the EmotiW 2020 AV Group-Level Emotion sub-challenge shows a final test accuracy of 65.74% for the multimodal approach, approximately 18% higher than the official baseline. The results show that using activity recognition pretraining offers performance advantages for group-emotion recognition and that audio is essential to improve the accuracy and robustness of video-based recognition.
Morphing attacks are one of the many threats that are constantly affecting deep face recognition systems. It consists of selecting two faces from different individuals and fusing them into a final image that contains the identity information of both. In this work, we propose a novel regularisation term that takes into account the existent identity information in both and promotes the creation of two orthogonal latent vectors. We evaluate our proposed method (OrthoMAD) in five different types of morphing in the FRLL dataset and evaluate the performance of our model when trained on five distinct datasets. With a small ResNet-18 as the backbone, we achieve state-of-the-art results in the majority of the experiments, and competitive results in the others.
Biometric recognition and presentation attack detection (PAD) methods strongly rely on deep learning algorithms. Though often more accurate, these models operate as complex black boxes. Interpretability tools are now being used to delve deeper into the operation of these methods, which is why this work advocates their integration in the PAD scenario. Building upon previous work, a face PAD model based on convolutional neural networks was implemented and evaluated both through traditional PAD metrics and with interpretability tools. An evaluation on the stability of the explanations obtained from testing models with attacks known and unknown in the learning step is made. To overcome the limitations of direct comparison, a suitable representation of the explanations is constructed to quantify how much two explanations differ from each other. From the point of view of interpretability, the results obtained in intra and inter class comparisons led to the conclusion that the presence of more attacks during training has a positive effect in the generalisation and robustness of the models. This is an exploratory study that confirms the urge to establish new approaches in biometrics that incorporate interpretability tools. Moreover, there is a need for methodologies to assess and compare the quality of explanations.