1
|
Patel Y, Shah T, Dhar MK, Zhang T, Niezgoda J, Gopalakrishnan S, Yu Z. Integrated image and location analysis for wound classification: a deep learning approach. Sci Rep 2024; 14:7043. [PMID: 38528003 DOI: 10.1038/s41598-024-56626-w] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [MESH Headings] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 11/01/2023] [Accepted: 03/08/2024] [Indexed: 03/27/2024] Open
Abstract
The global burden of acute and chronic wounds presents a compelling case for enhancing wound classification methods, a vital step in diagnosing and determining optimal treatments. Recognizing this need, we introduce an innovative multi-modal network based on a deep convolutional neural network for categorizing wounds into four categories: diabetic, pressure, surgical, and venous ulcers. Our multi-modal network uses wound images and their corresponding body locations for more precise classification. A unique aspect of our methodology is incorporating a body map system that facilitates accurate wound location tagging, improving upon traditional wound image classification techniques. A distinctive feature of our approach is the integration of models such as VGG16, ResNet152, and EfficientNet within a novel architecture. This architecture includes elements like spatial and channel-wise Squeeze-and-Excitation modules, Axial Attention, and an Adaptive Gated Multi-Layer Perceptron, providing a robust foundation for classification. Our multi-modal network was trained and evaluated on two distinct datasets comprising relevant images and corresponding location information. Notably, our proposed network outperformed traditional methods, reaching an accuracy range of 74.79-100% for Region of Interest (ROI) without location classifications, 73.98-100% for ROI with location classifications, and 78.10-100% for whole image classifications. This marks a significant enhancement over previously reported performance metrics in the literature. Our results indicate the potential of our multi-modal network as an effective decision-support tool for wound image classification, paving the way for its application in various clinical contexts.
Collapse
Affiliation(s)
- Yash Patel
- Department of Computer Science, University of Wisconsin-Milwaukee, Milwaukee, WI, USA
| | - Tirth Shah
- Department of Computer Science, University of Wisconsin-Milwaukee, Milwaukee, WI, USA
| | - Mrinal Kanti Dhar
- Department of Computer Science, University of Wisconsin-Milwaukee, Milwaukee, WI, USA
| | - Taiyu Zhang
- Department of Computer Science, University of Wisconsin-Milwaukee, Milwaukee, WI, USA
| | - Jeffrey Niezgoda
- Advancing the Zenith of Healthcare (AZH) Wound and Vascular Center, Milwaukee, WI, USA
| | | | - Zeyun Yu
- Department of Computer Science, University of Wisconsin-Milwaukee, Milwaukee, WI, USA.
- Department of Biomedical Engineering, University of Wisconsin-Milwaukee, Milwaukee, WI, USA.
| |
Collapse
|
2
|
Anisuzzaman DM, Patel Y, Rostami B, Niezgoda J, Gopalakrishnan S, Yu Z. Multi-modal wound classification using wound image and location by deep neural network. Sci Rep 2022; 12:20057. [PMID: 36414660 PMCID: PMC9681740 DOI: 10.1038/s41598-022-21813-0] [Citation(s) in RCA: 4] [Impact Index Per Article: 2.0] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Received: 04/26/2022] [Accepted: 10/04/2022] [Indexed: 11/24/2022] Open
Abstract
Wound classification is an essential step of wound diagnosis. An efficient classifier can assist wound specialists in classifying wound types with less financial and time costs and help them decide on an optimal treatment procedure. This study developed a deep neural network-based multi-modal classifier using wound images and their corresponding locations to categorize them into multiple classes, including diabetic, pressure, surgical, and venous ulcers. A body map was also developed to prepare the location data, which can help wound specialists tag wound locations more efficiently. Three datasets containing images and their corresponding location information were designed with the help of wound specialists. The multi-modal network was developed by concatenating the image-based and location-based classifier outputs with other modifications. The maximum accuracy on mixed-class classifications (containing background and normal skin) varies from 82.48 to 100% in different experiments. The maximum accuracy on wound-class classifications (containing only diabetic, pressure, surgical, and venous) varies from 72.95 to 97.12% in various experiments. The proposed multi-modal network also showed a significant improvement in results from the previous works of literature.
Collapse
Affiliation(s)
- D. M. Anisuzzaman
- grid.267468.90000 0001 0695 7223Department of Computer Science, University of Wisconsin-Milwaukee, Milwaukee, WI USA
| | - Yash Patel
- grid.267468.90000 0001 0695 7223Department of Computer Science, University of Wisconsin-Milwaukee, Milwaukee, WI USA
| | - Behrouz Rostami
- grid.267468.90000 0001 0695 7223Department of Electrical Engineering, University of Wisconsin-Milwaukee, Milwaukee, WI 53211 USA
| | - Jeffrey Niezgoda
- Advancing the Zenith of Healthcare (AZH) Wound and Vascular Center, Milwaukee, WI USA
| | - Sandeep Gopalakrishnan
- grid.267468.90000 0001 0695 7223College of Nursing, University of Wisconsin Milwaukee, Milwaukee, WI USA
| | - Zeyun Yu
- grid.267468.90000 0001 0695 7223Department of Computer Science, University of Wisconsin-Milwaukee, Milwaukee, WI USA ,grid.267468.90000 0001 0695 7223Big Data Analytics and Visualization Laboratory, Department of Biomedical Engineering, University of Wisconsin-Milwaukee, 3200 N. Cramer St, EMS E327, Milwaukee, WI 53211 USA
| |
Collapse
|
3
|
Montull L, Slapšinskaitė-Dackevičienė A, Kiely J, Hristovski R, Balagué N. Integrative Proposals of Sports Monitoring: Subjective Outperforms Objective Monitoring. SPORTS MEDICINE - OPEN 2022; 8:41. [PMID: 35348932 PMCID: PMC8964908 DOI: 10.1186/s40798-022-00432-z] [Citation(s) in RCA: 1] [Impact Index Per Article: 0.5] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Download PDF] [Figures] [Subscribe] [Scholar Register] [Received: 12/07/2021] [Accepted: 02/26/2022] [Indexed: 12/20/2022]
Abstract
Current trends in sports monitoring are characterized by the massive collection of tech-based biomechanical, physiological and performance data, integrated through mathematical algorithms. However, the application of algorithms, predicated on mechanistic assumptions of how athletes operate, cannot capture, assess and adequately promote athletes' health and performance. The objective of this paper is to reorient the current integrative proposals of sports monitoring by re-conceptualizing athletes as complex adaptive systems (CAS). CAS contain higher-order perceptual units that provide continuous and multilevel integrated information about performer-environment interactions. Such integrative properties offer exceptional possibilities of subjective monitoring for outperforming any objective monitoring system. Future research should investigate how to enhance this human potential to contribute further to athletes' health and performance. This line of argument is not intended to advocate for the elimination of objective assessments, but to highlight the integrative possibilities of subjective monitoring.
Collapse
Affiliation(s)
- Lluc Montull
- Complex Systems in Sport Research Group, Institut Nacional d'Educació Física de Catalunya (INEFC), Universitat de Barcelona, Barcelona, Spain
- University School of Health and Sport, University of Girona, Girona, Spain
| | - Agne Slapšinskaitė-Dackevičienė
- Complex Systems in Sport Research Group, Institut Nacional d'Educació Física de Catalunya (INEFC), Universitat de Barcelona, Barcelona, Spain
- Department of Sports Medicine, Faculty of Nursing and Faculty of Public Health, Health Research Institute, Medical Academy, Lithuanian University of Health Sciences, Kaunas, Lithuania
| | - John Kiely
- Institute of Coaching and Performance, School of Sport and Wellbeing, University of Central Lancashire, Preston, PR1 2HE, UK
| | - Robert Hristovski
- Complex Systems in Sport Research Group, Faculty of Physical Education, Sport and Health, Ss. Cyril and Methodius University in Skopje, Skopje, Macedonia
| | - Natàlia Balagué
- Complex Systems in Sport Research Group, Institut Nacional d'Educació Física de Catalunya (INEFC), Universitat de Barcelona, Barcelona, Spain.
| |
Collapse
|
4
|
Flow dynamics during incremental velocity running. JOURNAL OF COMPLEXITY IN HEALTH SCIENCES 2021. [DOI: 10.21595/chs.2021.21858] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Indexed: 11/18/2022] Open
|
5
|
O'Connor PJ. Pain During a Marathon Run: Prevalence and Correlates in a Cross-Sectional Study of 1,251 Recreational Runners in 251 Marathons. Front Sports Act Living 2021; 3:630584. [PMID: 33644755 PMCID: PMC7902858 DOI: 10.3389/fspor.2021.630584] [Citation(s) in RCA: 3] [Impact Index Per Article: 1.0] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Download PDF] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Received: 11/17/2020] [Accepted: 01/11/2021] [Indexed: 11/26/2022] Open
Abstract
This cross-sectional study aimed to obtain fundamental knowledge about pain during a marathon run. During the first seven months of 2007, announcements on websites of interest to marathon runners were used to recruit participants. A total of 1,251 runners (550 female runners) completed a 41-question online survey about the location and intensity of their primary pain during their last marathon and potentially related variables [perception of effort during the marathon, number of prior marathons run, typical pain intensity during training runs, percentage of training days with run-induced pain, highest intensity pain ever experienced]. Pain location was selected from a list of 27 specified body sites covering the entire body. Kilometer at which pain first occurred indexed pain threshold. Pain intensity at the primary location of pain was measured with a standardized, well-validated 0–10 pain intensity scale. Pearson correlations and multiple regression quantified the associations between average pain intensity and other variables. Sex-related differences in pain were tested using independent t-tests. Effort ratings (6–20) were added as a covariate in an ANCOVA to test if perceived effort accounted for possible sex-related differences in pain. Based on the available research, it was hypothesized that: (i) most runners would report moderate intensity pain, (ii) pain would be associated with both exercise intensity during the marathon and pain during training, and (iii) after adjusting for expected sex-related differences in perceived effort, females would experience pain earlier and rate the pain intensity as higher. All but two runners (99.8%) reported pain during a marathon, and most frequently in the anterior/medial thigh (17.1%), hamstring (10%), and calf (9.3%) locations. Pain threshold occurred at 25.3 ± 9.8 km (15.7 ± 6.1 miles) and the overall pain intensity of the run was 5.26 ± 2.45. No sex-related pain differences were found. Overall pain intensity during a marathon was significantly associated with: pain intensity during training runs (r = 0.39), percentage of training days with run-induced pain (r = 0.23), highest intensity pain ever experienced (r = 0.23), number of prior marathons (r = −0.18), and intensity of effort (r = 0.11) (all P < 0.001). Most runners experience moderate to very strong intensity pain during a marathon; the pain was independent of biological sex, and the pain is weakly associated with marathon race experience, pain during training, race effort, and the highest intensity of pain ever experienced.
Collapse
Affiliation(s)
- Patrick J O'Connor
- Department of Kinesiology, University of Georgia, Athens, GA, United States
| |
Collapse
|
6
|
Afraimovich VS, Zaks MA, Rabinovich MI. Mind-to-mind heteroclinic coordination: Model of sequential episodic memory initiation. CHAOS (WOODBURY, N.Y.) 2018; 28:053107. [PMID: 29857651 DOI: 10.1063/1.5023692] [Citation(s) in RCA: 1] [Impact Index Per Article: 0.2] [Reference Citation Analysis] [Abstract] [MESH Headings] [Track Full Text] [Subscribe] [Scholar Register] [Indexed: 06/08/2023]
Abstract
Retrieval of episodic memory is a dynamical process in the large scale brain networks. In social groups, the neural patterns, associated with specific events directly experienced by single members, are encoded, recalled, and shared by all participants. Here, we construct and study the dynamical model for the formation and maintaining of episodic memory in small ensembles of interacting minds. We prove that the unconventional dynamical attractor of this process-the nonsmooth heteroclinic torus-is structurally stable within the Lotka-Volterra-like sets of equations. Dynamics on this torus combines the absence of chaos with asymptotic instability of every separate trajectory; its adequate quantitative characteristics are length-related Lyapunov exponents. Variation of the coupling strength between the participants results in different types of sequential switching between metastable states; we interpret them as stages in formation and modification of the episodic memory.
Collapse
Affiliation(s)
- V S Afraimovich
- Instituto de Investigación en Comunicación Óptica, Universidad Autónoma de San Luis Potosí, 78220 San Luis Potosí, Mexico
| | - M A Zaks
- Institute of Physics, Humboldt University of Berlin, 12489 Berlin, Germany
| | - M I Rabinovich
- BioCircuits Institute, University of California, San Diego, La Jolla, California 92093-0328, USA
| |
Collapse
|