In this paper we propose a head detection method using range data from a stereo camera. The method is based on a technique that has been introduced in the domain of voxel data. For application in stereo cameras, the technique is extended (1) to be applicable to stereo data, and (2) to be robust with regard to noise and variation in environmental settings. The method consists of foreground selection, head detection, and blob separation, and, to improve results in case of misdetections, incorporates a means for people tracking. It is tested in experiments with actual stereo data, gathered from three distinct real-life scenarios. Experimental results show that the proposed method performs well in terms of both precision and recall. In addition, the method was shown to perform well in highly crowded situations. From our results, we may conclude that the proposed method provides a strong basis for head detection in applications that utilise stereo cameras.
MULTIFILE
In this paper, we address the problem of people detection and tracking in crowded scenes using range cameras. We propose a new method for people detection and localisation based on the combination of background modelling and template matching. The method uses an adaptive background model in the range domain to characterise the scene without people. Then a 3D template is placed in possible people locations by projecting it in the background to reconstruct a range image that is most similar to the observed range image. We tested the method on a challenging outdoor dataset and compared it to two methods that each shares one characteristic with the proposed method: a similar template-based method that works in 2D and a well-known baseline method that works in the range domain. Our method performs significantly better, does not deteriorate in crowded environments and runs in real time.
In recent years, drones have increasingly supported First Responders (FRs) in monitoring incidents and providing additional information. However, analysing drone footage is time-intensive and cognitively demanding. In this research, we investigate the use of AI models for the detection of humans in drone footage to aid FRs in tasks such as locating victims. Detecting small-scale objects, particularly humans from high altitudes, poses a challenge for AI systems. We present first steps of introducing and evaluating a series of YOLOv8 Convolutional Neural Networks (CNNs) for human detection from drone images. The models are fine-tuned on a created drone image dataset of the Dutch Fire Services and were able to achieve a 53.1% F1-Score, identifying 439 out of 825 humans in the test dataset. These preliminary findings, validated by an incident commander, highlight the promising utility of these models. Ongoing efforts aim to further refine the models and explore additional technologies.
MULTIFILE
Over a million people in the Netherlands have type 2 diabetes (T2D), which is strongly related to overweight, and many more people are at-risk. A carbohydrate-rich diet and insufficient physical activity play a crucial role in these developments. It is essential to prevent T2D, because this condition is associated with a reduced quality of life, high healthcare costs and premature death due to cardiovascular diseases. The hormone insulin plays a major role in this. This hormone lowers the blood glucose concentration through uptake in body cells. If an excess of glucose is constantly offered, initially the body maintains blood glucose concentration within normal range by releasing higher concentrations of insulin into the blood, a condition that is described as “prediabetes”. In a process of several years, this compensating mechanism will eventually fail: the blood glucose concentration increases resulting in T2D. In the current healthcare practice, T2D is actually diagnosed by recognizing only elevated blood glucose concentrations, being insufficient for identification of people who have prediabetes and are at-risk to develop T2D. Although the increased insulin concentrations at normal glucose concentrations offer an opportunity for early identification/screening of people with prediabetes, there is a lack of effective and reliable methods/devices to adequately measure insulin concentrations. An integrated approach has been chosen for identification of people at-risk by using a prediabetes screening method based on insulin detection. Users and other stakeholders will be involved in the development and implementation process from the start of the project. A portable and easy-to-use demonstrator will be realised, based on rapid lateral flow tests (LFTs), which is able to measure insulin in clinically relevant samples (serum/blood) quickly and reliably. Furthermore, in collaboration with healthcare professionals, we will investigate how this screening method can be implemented in practice to contribute to a healthier lifestyle and prevent T2D.
In the Netherlands, 125 people suffer a stroke every day, which annually results in 46.000 new stroke patients Stroke patients are confronted with combinations of physical, psychological and social consequences impacting their long term functioning and quality of live. Fortunately many patients recover to their pre-stroke level of functioning, however, almost half of them never will. Consequently, rehabilitation often means that patients need to adapt to a new reality in their lives, requiring not only physical but also psychosocial adjustments. Nurses play a key role during rehabilitation of stroke patients. However, when confronted with psychosocial problems, they often feel insecure about identifying the specific psycho-social needs of the individual patient and providing adequate care. In our project ‘Early Detection of Post-Stroke Depression’, (SIA RAAK; 2010-12-36P), we developed a toolkit focusing on early identification of depression after stroke continued with interventions nurses can use during hospitalisation. During this project it became clear that evidence regarding possible interventions is scarce and inclusive. Moreover feasibility of interventions is often not confirmed. Our project showed that during the period of hospital admission patients and health care providers strongly focus on surviving the stroke and on the physical rehabilitation. Therefore, we concluded that to make one step beyond we first have to go one step back. To strengthen psychosocial care for patients after stroke we have to add, reconsider and shape knowledge in context of health care practices in a systematic way, resulting in evidence based and practice informed stepping stones. With this project we aim to collect these stepping stones and develop a nursing care programme that improves psychosocial well-being of patients after stroke, is tailored to the particular concerns and needs of patients, and is considered feasible for use in the usual care process of nurses in the stroke rehabilitation pathway.
Antimicrobial Resistance (AMR), the ability of micro-organisms to resist antibiotics, is associated with ~4.9 million deaths globally, reported in 2022. In the EU alone, more than 35.000 people die from antimicrobial-resistant infections annually, resulting in loss of life as well as €1.5Bn/year in healthcare costs and productivity losses. Rapid diagnostics tests are needed, current testing takes between 24 hours to a few days (for slow growing microorganisms), delaying patient treatment and severely impacting treatment outcomes. SoundCell BV have developed a technique (TRL5), for real-time detection of bacteria's viability in the presence of antibiotics. Nano-mechanical vibration of an ultrathin graphene sheet correlates to viability of bacteria immobilized on this sheet. Bacterial motion is transferred to this sheet, and movement of this sheet is tracked via a high-speed laser. Living bacteria produce a strong signal, which diminishes when antibiotics kill them. Unaffected by growth rates, results are achieved in one hour with this technique. This technology opens up possibility for rapid diagnostics of antibiotic resistance in patients with infections of slow growing pathogens (such as mycobacteria and yeast). In such cases the time to result is slowest, significantly delaying effective patient treatment. We aim to validate this technique in our clinical microbiology laboratory.
