Open Research Oklahoma (Oklahoma State Univ.)
Not a member yet
42035 research outputs found
Sort by
Quality improvement project to improve provider education of helping babies breathe in Malawi, Africa
No full textIntroduction: Neonatal mortality rates are highest in developing countries, primarily due to a lack of healthcare. This was witnessed first-hand during a global health trip to Malawi, Africa in 2023. One such way to reduce neonatal mortality rates is the teaching and implementation of quality neonatal resuscitation in healthcare facilities. In response to high worldwide neonatal mortality rates, the American Academy of Pediatrics (AAP) created the Neonatal Resuscitation Program (NRP). Subsequently, the Helping Babies Breathe (HBB) curriculum, a subset of NRP, has been shown to reduce neonatal mortality rates in developing countries where resources are further limited. The goal of this project is to provide HBB training resources and education to hospital staff members at Child Legacy International (CLI).Methods: A survey was sent out to the physicians and staff members at Child Legacy International to assess current knowledge on neonatal resuscitation. A video detailing how to perform and implement Helping Babies Breathe will be recorded and sent to the staff to view. Researchers will be available to answer any questions or go over any specific topics for staff members as needed. Additionally, handouts and informational brochures will also be provided to each staff member who completes the survey. After 3 months' time, the same survey will be completed again by staff members to compare and contrast knowledge improvement. Results will be analyzed to see if the videos and handouts provided helped improve knowledge of NRP in staff members at CLI in Malawi.Results: Results for this project have not yet been collected.Conclusions: Thorough implementation of HBB, a subset of NRP, is proven to reduce the mortality rates of neonates. Based on first hand experience from working at CLI in Malawi, Africa, the staff members knowledge of NRP was lacking. It is expected that with proper training and resources, the knowledge and understanding of when and how to use HBB on neonates will improve with this quality improvement project. And with improved knowledge and implementation of HBB, neonatal mortality rates should also decrease as has been shown worldwide by the AAP. Results and conclusions from this project are currently pending
Computed tomography visualization of wormian bones in the pen-tailed tree shrew (Ptilocercus lowii)
No full textWormian bones are variably ossifying bones within the sutures of skulls. While widespread among mammals, they do not occur in every specimen of species that are known to form Wormian bones. While the exact reason(s) for why these bones form is unknown, it has been hypothesized that they may form due to prolonged cranial stress. Wormian bones in primates have been studied for over a century, but to our knowledge, they have not yet been reported from Ptilocercus lowii (Pen-tailed tree shrew), a member of Scandentia, a mammalian order sometimes recovered as an outgroup to Primatomorpha, the group that includes primates. We focused on two main questions: 1) How does the location and morphology of Wormian bones vary within P. lowii? and 2) Are the locations of the Wormian bones in P. lowii the same as in primates?To examine tree shrew skulls for Wormian bones, we sourced computed tomography (CT) scans from Morphosource. The Ptilocercus lowii skulls are female specimens from different nature reserves in Malaysia. These scans were first viewed on surface meshes to search for Wormian bones. The CT scans with potential Wormian bones were then uploaded to Avizo 2020.1. Once in Avizo, modifications were made to the scan to better locate the Wormian bone. Avizo’s transformation function was used to align the scan into anatomical position. The gamma filter was then applied to show only the whitest pixels. This technique ensured that only bone was visible and no debris from the scanning process was picked up in the image. Each slice of the CT scan was viewed in search of the potential Wormian bone before segmenting it manually using the Avizo paintbrush. The bone was then examined by all authors, and once it was agreed that it was indeed a Wormian bone, portions of the adjacent bones were segmented manually.Both tree shrew specimens have Wormian bones, but they are not in the same location in the skull. Specimen USNM 481107 has a Wormian bone in the lambda (where the sagittal and lambdoidal sutures meet). This Wormian bone can also be classified as an interparietal. USNM 481103 has a Wormian bone in the suture between the frontal and nasal bones and lacks a bone in the lambda. Interestingly, both Wormian bones overlapped surrounding skull bones along parts of their sutures, a morphology not readily apparent from the external views alone. Wormian bones occur in similar locations in primates. These findings show that the locations of Wormian bones are variable within P. lowii; however, with only two specimens, it is difficult to draw firm conclusions about what this variation means. These data also show us that the locations are consistent among primates and at least one scandentian taxon, suggesting the possibility of a homologous developmental origin for these bones.We plan to expand our sample to determine whether additional specimens will show consistent Wormian bone locations within P. lowii and to further compare the varying locations of Wormian bones in P. lowii to those of other mammals
Enhancing robustness of convolution neural networks against adversarial attacks with approximate adders and multipliers
No full textThe utilization of deep learning models such as convolutional neural networks (CNNs) in real-time applications has been increasing. This surge in the deployment of CNNs in self-driving vehicles and image classification tasks has made them susceptible to adversarial attacks. Adversarial attacks involve deliberately designed disturbances or the injection of noise into image pixels, posing a threat to safety-critical systems and potentially resulting in fatal incidents in automotive settings. Current research efforts to combat these attacks primarily focus on software-based strategies such as adversarial training and weight quantization methods. However, these defenses are computationally intensive and can escalate hardware resource requirements. Given that CNNs can tolerate minor errors in their internal layers, we employed approximate computing to assess their resilience against adversarial attacks. In our study, we introduced the approximate mirror adder (AMA) topology to replace exact full adders in the floating-point multipliers of the convolution and linear layers within CNNs. This enabled the implementation of a 32-bit approximate floating-point multiplier. We conducted black-box adversarial attacks using the Fast Gradient Sign Method (FGSM) algorithm on both the precise and approximate versions of the LeNet-5 CNN classifier. Through a comparative analysis, we investigated how different approximate hardware components responded to these attacks and their impact on classifier output and accuracy. Additionally, we evaluated the potential energy savings, power efficiency, and area advantages that can be achieved through approximate computing. By pursuing a hardware-software co-design approach such as Neural Architecture Search (NAS), we anticipate enhancing the resilience of CNNs against adversarial attacks
Eastern wild turkey population dynamics and nest survival in a declining population
No full textThe decline in wild turkey populations, particularly in regions like the southeastern United States and Oklahoma, presents significant challenges for conservation efforts. This study delves into the complex dynamics influencing nest predation and reproductive success in wild turkey populations, with a specific focus on McCurtain County, Oklahoma. Through a comprehensive field study, our objective was to shed light on the factors underlying the observed declines and provide guidance for targeted management strategies. We conducted a detailed assessment of current reproductive and survival vital rates to determine the trajectory of the turkey population in McCurtain County, Oklahoma. By deriving vital rates directly from field observations, we aimed to identify the specific factors exerting the greatest influence on population growth rate. Additionally, we characterized nest sites and evaluated turkey nest survival to explore the potential relationships between nest site and landscape characteristics and nest success. Our findings reveal the intricate interplay of various factors contributing to the decline in wild turkey populations. Predation emerges as a significant driver of nest failure, highlighting the importance of understanding predator-prey dynamics in shaping reproductive success. Furthermore, landscape features may play a role in influencing nest predation risk, underscoring the need for comprehensive landscape-scale assessments.This study provides valuable insights for conservationists and wildlife managers seeking to address the decline in wild turkey populations. By understanding the drivers of population declines and identifying key factors affecting reproductive success, we can develop targeted management strategies to support turkey population recovery. Habitat restoration efforts, predator management, and landscape conservation measures emerge as potential avenues for mitigating the impacts of nest predation and promoting turkey population growth
4-H Home Demonstration Club manual: Fifth year
No full textThe Oklahoma Cooperative Extension Service periodically issues revisions to its publications. The most current edition is made available. For access to an earlier edition, if available for this title, please contact the Oklahoma State University Library Archives by email at [email protected] or by phone at 405-744-6311
Deep learning for fast quality assessment
No full textThis report presents an automated system for classifying the quality of Focused Assessment with Sonography for Trauma (FAST) ultrasound exams. Our dataset consisted of 441 FAST exams with 3,161 videos, graded from 1 (poor quality) to 5 (good quality). We deemed exams rated as 1 or 2 as poor quality, and those rated 3 to 5 as good quality. Our approach involved assigning the quality label of the exam to each video and then to each frame within, as frames lacked individual labels. We utilized a custom CNN autoencoder for image compression, and then used the encoder for further classification. The classification of a video or an exam as poor quality was determined if at least half of its frames or videos, respectively, were assessed as such. The encoder-classifier model outperformed the transfer learning model, with our ensemble testing accuracy achieving 98% for video quality and 100% sensitivity and specificity at the exam level. The performance of our encoder-classifier network significantly surpassed traditional transfer learning methods, demonstrating its effectiveness in accurately assessing ultrasound exam quality.
Deep learning methods have shown remarkable efficacy in the domain of medical imaging, offering promising advancements for various diagnostic procedures. Nonetheless, a significant barrier to their widespread adoption is the opacity of their decision- making processes. Several techniques exist to explain model decisions; some focus on patterns specific to individual predictions, while others are interested in the overall pattern or logic of the decision. In this paper, we present a new modification to existing local explainability techniques—DeepLIFT and Integrated Gradient. These modifications aim to enhance local explainability, providing clearer insights into the model’s decision-making process.
Furthermore, we introduce the Self-Organizing Map (SOM) as a novel approach to global explainability. This technique provides a global- level understanding by revealing patterns and trends in how deep learning models make classification decisions across a broad dataset. We combined these two new methods and applied them in a Focused Assessment with Sonography in trauma exam classification problem. Through our comprehensive analysis using local and global explainability techniques, we discovered that the clarity (sharpness) and informational content (density) of FAST ultrasound frames are pivotal attributes leveraged by deep learning algorithms for quality classification. This insight not only enhances our understanding of the underlying mechanisms of these models but also opens new avenues for improving the accuracy and reliability of image classification problems
Dung beetle communities and impact on dung degradation and horn fly population under different management practices
No full textDung beetles (Coleoptera: Scarabaeidae) are vital contributors to waste recycling in grazed ecosystems, particularly those involving large vertebrates such as cattle and bison. This study evaluated the diversity and abundance of dung beetle communities across various grazing systems in Oklahoma, with a focus on comparing cattle management to bison-grazed landscapes. Monthly samples with baited pitfall traps were collected from May to August 2024 across multiple sites, including organically managed ranches, conventional grazing systems, and the bison-grazed Tallgrass Prairie Preserve. A total of 42,049 dung beetles were collected comprising eleven species or species groups. Significant differences in community composition and abundance existed by month and grazing systems. Species diversity peaked in the spring and declined towards late summer, particularly in conventionally managed grazing systems. In systems with more dung beetles, dung decomposition was faster with 80% degradation in a 16-day period. During periods of peak dung beetle activity horn fly, Haematobia irritans, numbers were also lower despite organic management. These findings underscore the positive impact of organic and rotational grazing practices, which support higher dung beetle diversity compared to conventional systems. Furthermore, this research highlights the essential role of dung beetles in nutrient cycling and soil health, particularly in grazing landscapes that incorporate adaptive management practices. These results reinforce the ecological importance of dung beetles in maintaining pasture health, improving cattle production systems, and reducing the need for chemical fly control methods. By enhancing dung breakdown pest species such as horn flies that breed in manure can be reduced
Aeroacoustic characterization of coaxial ducted propellers/rotors for small UAVs
No full textThe growing interest in electric vertical takeoff and landing (eVTOL) aircraft has highlighted the need for a comprehensive understanding of the aerodynamic and acoustic characteristics of open versus ducted rotors. This numerical study investigates the aerodynamic performance and acoustic behavior of open and ducted coaxial counter-rotating propellers, with a focus on the effects of a duct and rotor axial spacing on the frequency spectrum and directivity of small UAV propeller noise. Using a high-fidelity numerical flow solver, charLES, which performs a compressible, wall-modeled Large Eddy Simulation (LES) capable of resolving turbulence and acoustics. A common UAV propeller with a diameter of 9.4-inches was modeled in the coaxial configuration, with variable spacing between the fore- and aft-rotor.Acustom designed duct was created. Far-field noise propagation is computed using the permeable Ffowcs-Williams & Hawkings technique. The results demonstrate that the ducted configuration exhibits a 4% increase in total thrust compared to the open rotors, attributable to reduced tip leakages and improved inflow. Increasing axial spacing between front and rear rotors leads to a slight increase in total predicted thrust due to reduced wake impingement on the rear rotor.
Acoustically, the duct effectively shields only the blade passing frequency (BPF, 200 Hz) tone, despite duct mode analysis indicating that the BPF, 3BPF, and 5BPF tones all become cut-off. At frequencies above 1,500 Hz, the ducted configuration generates higher broadband noise levels due to the interaction between the duct boundary layer and the rotor tip vortices with higher turbulent intensity. The ducted configuration exhibits higher overall sound pressure levels (OASPL) and a distinct two-lobe directivity pattern, while the open configuration has a smeared directivity with lower OASPL in the fore arc due to BPF tone dominance. The fundamental tone at 200 Hz forms a directivity with the peak near 100◦ indicating a rotor-only tone, regardless of the rotor spacing; in contrast, the first harmonic tone at 400 Hz depicts a distinctive two-lobe pattern indicating an equally-split interaction tone between the front- and rear-rotor harmonics, except for the 1.5-inch spacing case. Both of the tone intensities decrease with the rotor spacing. The rotor spacing predominantly affects low-frequency tones (200 − 1500 Hz), which decrease their amplitude with increasing spacing, while high-frequency noise (2500 − 4000 Hz) dominated by broadband shows a little sensitivity to the rotor spacing. This suggests that the rotor spacing primarily impacts on large-scale structures in the wakes rather than small-scale features, such as turbulence and tip vortices in the wall boundary layer. This study contributes to the understanding of the aeroacoustic characteristics of ducted coaxial counter-rotating propellers and provides insights into the effects of duct integration and rotor spacing on UAV propeller noise. The findings can inform the design and optimization of quiet and efficient eVTOL propulsion systems
Relationship of academic self-efficacy and mindset on academic resilience
No full textThis study explores the contributive factors of academic self-efficacy and mindset upon academic resilience in a population of 258 BSN nursing students in the Southwest region of the U.S. A confirmatory factor analysis was the first step to identifying latent variables that are supported in the research. The measurement models for the ARS-30, SELF-A, and mindset tools demonstrated adequate model fit indices, including the chi-square model of fit, RMSEA, and CFI. The factor loadings were evaluated for adequate fit. Three latent factors of academic self-efficacy were identified studying, test-prep, and note-taking. The latent factors of academic resilience were identified as perseverance, motivation, and self-regulation. The structural equation model (SEM) for academic resilience met the recommendations for composite reliability. The SEM analysis supports that mindset and academic self-efficacy are significantly associated with academic resilience.
The amount of variance of academic resilience explained by academic mindset was r²=0.124 and for academic self-efficacy r²=0.432. When regressing academic resilience on academic self-efficacy and academic mindset the r2=0.456, which supports that academic mindset and academic self-efficacy share some of the explained variance of academic resilience. The correlation of mindset with academic self-efficacy was found to be 0.335. The analysis supports the hypothesis “mindset and academic self-efficacy are significantly associated with academic resilience.”
The mixed method of concurrent triangulation design and the analysis of quantitative and qualitative data provided insight into research questions one, two, and three. The variance in academic resilience is associated with both academic self-efficacy and academic mindset. Academic self-efficacy was shown to have a higher association with academic resilience than mindset in this population than mindset
Essays on corporate culture
No full textEssay one examines the influence of corporate culture on mergers and acquisitions (M&A) outcomes using text data from sections 1 and 7 of 10-K filings and a culture dictionary. Through ordinary least squares (OLS) regression models and an event study, the research reveals how cultural dimensions like innovation, quality, integrity, teamwork, and respect affect M&A outcomes. Key findings include that the aggregate culture score of acquiring firms does not significantly influence their announcement returns, but the specific cultures of respect and integrity do. Combined culture scores of both acquirers and targets significantly impact the announcement returns of acquirers. While the aggregate culture score of acquiring firms does not significantly affect long-term financial performance, individual scores for innovation and quality enhance it. The aggregate culture score of target firms significantly impacts deal premiums, with higher integrity scores raising premiums. The culture scores of acquiring firms do not significantly impact deal premiums. Higher aggregate culture scores for both acquiring and target firms, especially in innovation, quality, and integrity, expedite deal completion, fostering trust and efficiency in the M&A process. However, teamwork does not significantly affect M&A outcomes.Essay two analyzes how the cultural traits of US corporate bond-issuing companies influence their credit ratings and bond yields. The study combines qualitative corporate culture assessments with a culture dictionary to derive scores for cultural traits. It then uses regression models to investigate the relationship between cultural dimensions and bonds’ ratings and yields. The study finds that the aggregate cultural scores of the bond-issuing firms, particularly innovation and quality, are significantly correlated with higher credit ratings and lower bond yields, indicating that investors perceive these traits as indicators of lower risk and better management quality. The paper provides empirical evidence that corporate culture is priced in the bond market and contributes to the existing corporate bond literature. It also sets a foundation for future research to explore the broader effects of various cultural dimensions of firms