Abstract
Biofeedback is a promising tool to complement conventional physical therapy by fostering active participation of neurologically impaired patients during treatment. This work aims at a user-centered design and usability assessment for different age groups of a novel wearable augmented reality application composed of a multimodal sensor network and corresponding control strategies for personalized biofeedback during gait training. The proposed solution includes wearable AR glasses that deliver visual cues controlled in real-time according to mediolateral center of mass position, sagittal ankle angle, or tibialis anterior muscle activity from inertial and EMG sensors. Control strategies include positive and negative reinforcement conditions and are based on the user's performance by comparing real-time sensor data with an automatically user-personalized threshold. The proposed solution allows ambulatory practice on daily scenarios, physiotherapists' involvement through a laptop screen, and contributes to further benchmark biofeedback regarding the type of sensor. Although old healthy adults with low academic degrees have a preference for guidance from an expert person, excellent usability scores (SUS scores: 81.25-96.87) were achieved with young and middle-aged healthy adults and one neurologically impaired patient.

Abstract
Underwater wireless communications are crucial for supporting multiple maritime activities, such as environmental monitoring and offshore wind farms. However, the challenging underwater environment continues to pose obstacles to the development of long-range, broadband underwater wireless communication systems. State of the art solutions are limited to long range, narrowband acoustics and short range, broadband radio or optical communications. This precludes real-time wireless transmission of imagery over long distancesIn this paper, we propose SAGE, a semantic-oriented underwater communications approach to enable real-time wireless imagery transmission over noisy and narrowband channels. SAGE extracts semantically relevant information from images at the sender located underwater and generates a text description that is transmitted to the receiver at the surface, which in turn generates an image from the received text description. SAGE is evaluated using BLIP for image-to-text and Stable Diffusion for text-to-image, showing promising image similarity between the original and the generated images, and a significant reduction in latency up to a hundred-fold, encouraging further research in this area. © 2024 IFIP.

Abstract
Post-hoc explanation methods have often been criticised for abstracting away the decision-making process of deep neural networks. In this work, we would like to provide natural language descriptions for what different layers of a vision backbone have learned. Our DeViL method generates textual descriptions of visual features at different layers of the network as well as highlights the attribution locations of learned concepts. We train a transformer network to translate individual image features of any vision layer into a prompt that a separate off-the-shelf language model decodes into natural language. By employing dropout both per-layer and per-spatial-location, our model can generalize training on image-text pairs to generate localized explanations. As it uses a pre-trained language model, our approach is fast to train and can be applied to any vision backbone. Moreover, DeViL can create open-vocabulary attribution maps corresponding to words or phrases even outside the training scope of the vision model. We demonstrate that DeViL generates textual descriptions relevant to the image content on CC3M, surpassing previous lightweight captioning models and attribution maps, uncovering the learned concepts of the vision backbone. Further, we analyze fine-grained descriptions of layers as well as specific spatial locations and show that DeViL outperforms the current state-of-the-art on the neuron-wise descriptions of the MILANNOTATIONS dataset.

Abstract
Considering the profound transformation affecting pathology practice, we aimed to develop a scalable artificial intelligence (AI) system to diagnose colorectal cancer from whole-slide images (WSI). For this, we propose a deep learning (DL) system that learns from weak labels, a sampling strategy that reduces the number of training samples by a factor of six without compromising performance, an approach to leverage a small subset of fully annotated samples, and a prototype with explainable predictions, active learning features and parallelisation. Noting some problems in the literature, this study is conducted with one of the largest WSI colorectal samples dataset with approximately 10,500 WSIs. Of these samples, 900 are testing samples. Furthermore, the robustness of the proposed method is assessed with two additional external datasets (TCGA and PAIP) and a dataset of samples collected directly from the proposed prototype. Our proposed method predicts, for the patch-based tiles, a class based on the severity of the dysplasia and uses that information to classify the whole slide. It is trained with an interpretable mixed-supervision scheme to leverage the domain knowledge introduced by pathologists through spatial annotations. The mixed-supervision scheme allowed for an intelligent sampling strategy effectively evaluated in several different scenarios without compromising the performance. On the internal dataset, the method shows an accuracy of 93.44% and a sensitivity between positive (low-grade and high-grade dysplasia) and non-neoplastic samples of 0.996. On the external test samples varied with TCGA being the most challenging dataset with an overall accuracy of 84.91% and a sensitivity of 0.996.

Abstract
Brain stroke, or a cerebrovascular accident, is a devastating medical condition that disrupts the blood supply to the brain, depriving it of oxygen and nutrients. Each year, according to the World Health Organization, 15 million people worldwide experience a stroke. This results in approximately 5 million deaths and another 5 million individuals suffering permanent disabilities. The complex interplay of various risk factors highlights the urgent need for sophisticated analytical methods to more accurately predict stroke risks and manage their outcomes. Machine learning and deep learning technologies offer promising solutions by analyzing extensive datasets including patient demographics, health records, and lifestyle choices to uncover patterns and predictors not easily discernible by humans. These technologies enable advanced data processing, analysis, and fusion techniques for a comprehensive health assessment. We conducted a comprehensive review of 25 review papers published between 2020 and 2024 on machine learning and deep learning applications in brain stroke diagnosis, focusing on classification, segmentation, and object detection. Furthermore, all these reviews explore the performance evaluation and validation of advanced sensor systems in these areas, enhancing predictive health monitoring and personalized care recommendations. Moreover, we also provide a collection of the most relevant datasets used in brain stroke analysis. The selection of the papers was conducted according to PRISMA guidelines. Furthermore, this review critically examines each domain, identifies current challenges, and proposes future research directions, emphasizing the potential of AI methods in transforming health monitoring and patient care.

Abstract
With the objective of developing new methods to acquire diagnostic information, the reconstruction of the broadband absorption coefficient spectra (mu a[lambda]) of healthy and chromophobe renal cell carcinoma kidney tissues was performed. By performing a weighted sum of the absorption spectra of proteins, DNA, oxygenated, and deoxygenated hemoglobin, lipids, water, melanin, and lipofuscin, it was possible to obtain a good match of the experimental mu a(lambda) of both kidney conditions. The weights used in those reconstructions were estimated using the least squares method, and assuming a total water content of 77% in both kidney tissues, it was possible to calculate the concentrations of the other tissue components. It has been shown that with the development of cancer, the concentrations of proteins, DNA, oxygenated hemoglobin, lipids, and lipofuscin increase, and the concentration of melanin decreases. Future studies based on minimally invasive spectral measurements will allow cancer diagnosis using the proposed approach.