scholarly journals WildGait: Learning Gait Representations from Raw Surveillance Streams

Sensors ◽  
2021 ◽  
Vol 21 (24) ◽  
pp. 8387
Author(s):  
Adrian Cosma ◽  
Ion Emilian Radoi
Keyword(s):  
Real World ◽  
Gait Recognition ◽  
Convolutional Network ◽  
Privacy Concerns ◽  
Learning Framework ◽  
Non Invasive ◽  
Current State ◽  
Single Person ◽  
Straight Line ◽  
Spatio Temporal

The use of gait for person identification has important advantages such as being non-invasive, unobtrusive, not requiring cooperation and being less likely to be obscured compared to other biometrics. Existing methods for gait recognition require cooperative gait scenarios, in which a single person is walking multiple times in a straight line in front of a camera. We address the challenges of real-world scenarios in which camera feeds capture multiple people, who in most cases pass in front of the camera only once. We address privacy concerns by using only motion information of walking individuals, with no identifiable appearance-based information. As such, we propose a self-supervised learning framework, WildGait, which consists of pre-training a Spatio-Temporal Graph Convolutional Network on a large number of automatically annotated skeleton sequences obtained from raw, real-world surveillance streams to learn useful gait signatures. We collected and compiled the largest pretraining dataset to date of anonymized walking skeletons called Uncooperative Wild Gait, containing over 38k tracklets of anonymized walking 2D skeletons. We make the dataset available to the research community. Our results surpass the current state-of-the-art pose-based gait recognition solutions. Our proposed method is reliable in training gait recognition methods in unconstrained environments, especially in settings with scarce amounts of annotated data.

scholarly journals Photoacoustic image-guided interventions

2019 ◽  
Vol 245 (4) ◽  
pp. 330-341 ◽  
Author(s):  
Madhumithra S Karthikesh ◽  
Xinmai Yang
Keyword(s):  
Drug Delivery ◽  
Real Time ◽  
Real World ◽  
Photoacoustic Imaging ◽  
Cost Effective ◽  
Hybrid Nanoparticles ◽  
Future Research ◽  
Non Invasive ◽  
Image Guided ◽  
Current State

Photoacoustic imaging has demonstrated its potential for diagnosis over the last few decades. In recent years, its unique imaging capabilities, such as detecting structural, functional and molecular information in deep regions with optical contrast and ultrasound resolution, have opened up many opportunities for photoacoustic imaging to be used during image-guided interventions. Numerous studies have investigated the capability of photoacoustic imaging to guide various interventions such as drug delivery, therapies, surgeries, and biopsies. These studies have demonstrated that photoacoustic imaging can guide these interventions effectively and non-invasively in real-time. In this minireview, we will elucidate the potential of photoacoustic imaging in guiding active and passive drug deliveries, photothermal therapy, and other surgeries and therapies using endogenous and exogenous contrast agents including organic, inorganic, and hybrid nanoparticles, as well as needle-based biopsy procedures. The advantages of photoacoustic imaging in guided interventions will be discussed. It will, therefore, show that photoacoustic imaging has great potential in real-time interventions due to its advantages over current imaging modalities like computed tomography, magnetic resonance imaging, and ultrasound imaging. Impact statement Photoacoustic imaging is an emerging modality for use in image-guided interventional procedures. This imaging technology has a unique ability to offer real-time, non-invasive, cost-effective, and radiation-free guidance in a real-world operating environment. This is substantiated in this article which sums up the current state and underlines promising results of research using photoacoustic imaging in guiding drug delivery, therapy, surgery, and biopsy. Hence, this minireview facilitates future research and real-world application of photoacoustic image-guided interventions.

Understanding Design Approaches and Evaluation Methods in mHealth applications Targeting Substance Use: Protocol for a Systematic Review (Preprint)

2022 ◽  
Author(s):  
Sahiti Kunchay ◽  
Ashley Linden-Carmichael ◽  
Stephanie Lanza ◽  
Saeed Abdullah
Keyword(s):  
Systematic Review ◽  
Substance Use ◽  
Real World ◽  
Large Scale ◽  
Final Analysis ◽  
Opioid Use ◽  
Privacy Concerns ◽  
Current State ◽  
Independent Evaluation ◽  
User Perspectives

BACKGROUND Substance use and use disorders in the US have had significant and devastating impacts on individuals and communities, and the escalating substance use crisis calls for urgent and innovative solutions to effectively detect and provide interventions for individuals in times of need. Recent mHealth-based approaches offer promising new opportunities to address these issues through ubiquitous devices. However, the design rationale, theoretical framing, and the mechanisms through which users' perspectives and experiences guide the design and deployment of such systems have not been analyzed in any prior systematic reviews. OBJECTIVE In this paper, we systematically review these approaches and applications for their feasibility, efficacy, and usability. Further, we evaluate whether human-centered research principles and techniques guide the design and development of these systems, and examine how the current state-of-art systems apply to real-world contexts. In an effort to gauge the applicability of these systems, we also investigate whether these approaches consider the effects of stigma and the privacy concerns related to collecting data on substance use. Lastly, we examine persistent challenges in the design and large-scale adoption of substance use intervention applications and draw inspiration from other domains of mHealth to suggest actionable reforms into the design and deployment of these applications. METHODS Four databases (PubMed, IEEE, JMIR and ACM DL) were searched over a five-year period (2016 - 2021) for articles evaluating connected mHealth approaches for substance use (alcohol use, marijuana use, opioid use, tobacco use, and substance co-use). Articles that will be included describe an mHealth detection or intervention targeting substance use and provided outcomes data and a discussion of design techniques and user perspectives. Independent evaluation will be conducted by one author, followed by secondary reviewer(s) who will check and validate themes and data. RESULTS This is a protocol for a systematic review, therefore results are not yet available. We are currently in the process of selecting the studies for inclusion in the final analysis. CONCLUSIONS To the best of our knowledge, this is the first systematic review to assess real-world applicability, scalability, and use of human-centered design and evaluation techniques in mHealth approaches targeting substance use. This study is expected to identify gaps in current substance use detection and intervention mHealth technologies and inform and motivate future development of such systems.

scholarly journals Differences in Cardiac Effects of Venoms from Tentacles and the Bell of Live Carukia barnesi: Using Non-Invasive Pulse Wave Doppler

Toxins ◽  
2020 ◽  
Vol 13 (1) ◽  
pp. 19
Author(s):  
Mark Little ◽  
Peter Pereira ◽  
Jamie Seymour
Keyword(s):  
Animal Models ◽  
Murine Model ◽  
Real World ◽  
Cardiac Dysfunction ◽  
Pulse Wave ◽  
Cardiac Effects ◽  
Post Exposure ◽  
Non Invasive ◽  
Cardiac Decompensation

Carukia barnesi was the first in an expanding list of cubozoan jellyfish whose sting was identified as causing Irukandji syndrome. Nematocysts present on both the bell and tentacles are known to produce localised stings, though their individual roles in Irukandji syndrome have remained speculative. This research examines differences through venom profiling and pulse wave Doppler in a murine model. The latter demonstrates marked measurable differences in cardiac parameters. The venom from tentacles (CBVt) resulted in cardiac decompensation and death in all mice at a mean of 40 min (95% CL: ± 11 min), whereas the venom from the bell (CBVb) did not produce any cardiac dysfunction nor death in mice at 60 min post-exposure. This difference is pronounced, and we propose that bell exposure is unlikely to be causative in severe Irukandji syndrome. To date, all previously published cubozoan venom research utilised parenterally administered venom in their animal models, with many acknowledging their questionable applicability to real-world envenomation. Our model used live cubozoans on anaesthetised mice to simulate normal envenomation mechanics and actual expressed venoms. Consequently, we provide validity to the parenteral methodology used by previous cubozoan venom research.

scholarly journals FetNet: a recurrent convolutional network for occlusion identification in fetoscopic videos

2020 ◽  
Vol 15 (5) ◽  
pp. 791-801 ◽  
Author(s):  
Sophia Bano ◽  
Francisco Vasconcelos ◽  
Emmanuel Vander Poorten ◽  
Tom Vercauteren ◽  
Sebastien Ourselin ◽  
...  
Keyword(s):  
Neural Network ◽  
Network Architecture ◽  
Laser Photocoagulation ◽  
Short Term Memory ◽  
Total Duration ◽  
Frame Rate ◽  
Superior Performance ◽  
Computer Assisted ◽  
Convolutional Network ◽  
Spatio Temporal

Abstract Purpose Fetoscopic laser photocoagulation is a minimally invasive surgery for the treatment of twin-to-twin transfusion syndrome (TTTS). By using a lens/fibre-optic scope, inserted into the amniotic cavity, the abnormal placental vascular anastomoses are identified and ablated to regulate blood flow to both fetuses. Limited field-of-view, occlusions due to fetus presence and low visibility make it difficult to identify all vascular anastomoses. Automatic computer-assisted techniques may provide better understanding of the anatomical structure during surgery for risk-free laser photocoagulation and may facilitate in improving mosaics from fetoscopic videos. Methods We propose FetNet, a combined convolutional neural network (CNN) and long short-term memory (LSTM) recurrent neural network architecture for the spatio-temporal identification of fetoscopic events. We adapt an existing CNN architecture for spatial feature extraction and integrated it with the LSTM network for end-to-end spatio-temporal inference. We introduce differential learning rates during the model training to effectively utilising the pre-trained CNN weights. This may support computer-assisted interventions (CAI) during fetoscopic laser photocoagulation. Results We perform quantitative evaluation of our method using 7 in vivo fetoscopic videos captured from different human TTTS cases. The total duration of these videos was 5551 s (138,780 frames). To test the robustness of the proposed approach, we perform 7-fold cross-validation where each video is treated as a hold-out or test set and training is performed using the remaining videos. Conclusion FetNet achieved superior performance compared to the existing CNN-based methods and provided improved inference because of the spatio-temporal information modelling. Online testing of FetNet, using a Tesla V100-DGXS-32GB GPU, achieved a frame rate of 114 fps. These results show that our method could potentially provide a real-time solution for CAI and automating occlusion and photocoagulation identification during fetoscopic procedures.

Spatio-Temporal Modeling on Village Land Feature Change

2012 ◽  
Vol 204-208 ◽  
pp. 2721-2725
Author(s):  
Hua Ji Zhu ◽  
Hua Rui Wu
Keyword(s):  
Real World ◽  
Temporal Change ◽  
End Users ◽  
Temporal Modeling ◽  
The Real ◽  
Change Type ◽  
Feature Change ◽  
Spatio Temporal ◽  
The Village ◽  
Land Changes

Village land continually changes in the real world. In order to keep the data up-to-date, data producers need update the data frequently. When the village land data are updated, the update information must be dispensed to the end-users to keep their client-databases current. In the real world, village land changes in many forms. Identifying the change type of village land (i.e. captures the semantics of change) and representing them in the data world can help end-users understand the change commonly and be convenient for end-users to integrate these change information into their databases. This work focuses on the model of the spatio-temporal change. A three-tuple model CAR for representing the spatio-temporal change is proposed based on the village land feature set before change and the village land feature set after change, change type and rules. In this model, the C denotes the change type. A denotes the attribute set; R denotes the judging rules of change type. The rule is described by the IF－THEN expressions. By the operations between R and A, the C is distinguished. This model overcomes the limitations of current methods. And more, the rules in this model can be easy realized in computer program.

scholarly journals Sequential convolutional network for behavioral pattern extraction in gait recognition

2021 ◽  
Author(s):  
Xinnan Ding ◽  
Kejun Wang ◽  
Chenhui Wang ◽  
Tianyi Lan ◽  
Liangliang Liu
Keyword(s):  
Gait Recognition ◽  
Behavioral Pattern ◽  
Pattern Extraction ◽  
Convolutional Network
Sign in / Sign up

Export Citation Format

Share Document