Guided reconstruction with conditioned diffusion models for unsupervised anomaly detection in brain MRIs.

The application of supervised models to clinical screening tasks is challenging due to the need for annotated data for each considered pathology. Unsupervised Anomaly Detection (UAD) is an alternative approach that aims to identify any anomaly as an outlier from a healthy training distribution. A prevalent strategy for UAD in brain MRI involves using generative models to learn the reconstruction of healthy brain anatomy for a given input image.

Domestic Use of E-Cargo Bikes and Other E-Micromobility: Protocol for a Multi-Centre, Mixed Methods Study.

Physical inactivity is a leading risk factor for non-communicable diseases. Climate change is now regarded as the biggest threat to global public health. Electric micromobility (e-micromobility, including e-bikes, e-cargo bikes, and e-scooters) has the potential to simultaneously increase people's overall physical activity while decreasing greenhouse gas emissions where it substitutes for motorised transport.

ExerG - an exergame-based training device for the rehabilitation of older adults: a functional model usability study.

Background: Exergames are interactive technology-based exercise programs. By combining physical and cognitive training components, they aim to preserve independence in older adults and reduce their risk of falling. This study explored whether primary end users (PEU, healthy older adults and patients with neurological and geriatric diagnoses) and secondary end users (SEU, health professionals) evaluated the ExerG functional model to be usable, providing a positive experience and therefore acceptable.

Realization of quasi-steady-state piled smouldering using sequential operation chambers for industrial treatment of biomass wastes.

Piled smouldering has great potential for treatment and utilization of biomass wastes. However, its unsteady-state nature limits its industrial utilization, as well as treatment of smoke. This article addresses this issue by proposing the sequential operation of numerous smouldering chambers to realize steady- or quasi-steady-state piled smouldering.

PolypNextLSTM: a lightweight and fast polyp video segmentation network using ConvNext and ConvLSTM.

Purpose: Commonly employed in polyp segmentation, single-image UNet architectures lack the temporal insight clinicians gain from video data in diagnosing polyps. To mirror clinical practices more faithfully, our proposed solution, PolypNextLSTM, leverages video-based deep learning, harnessing temporal information for superior segmentation performance with least parameter overhead, making it possibly suitable for edge devices.

Methods: PolypNextLSTM employs a UNet-like structure with ConvNext-Tiny as its backbone, strategically omitting the last two layers to reduce parameter overhead.