100 research outputs found
Patient Experience: Optimization of the Discharge Phase
No full textThis essay was written and conducted to dissect the discharge phase and discover ways to optimize the existing process, specifically at Maui Health. The focus of this essay will be on the current challenges during the discharge phase that affect the patients experience and ways I was able to optimize the discharge process. This essay will discuss the negative impacts of a poor patient experience on both the patient and the health system as a whole. Furthermore, I will go in depth on how I was able to leverage my skills and competencies to implement solutions to solve Maui Health’s ongoing issues such as the hospital visit folder, inpatient communication and post discharge call system
Virtual environment and its ground surface can influence locomotion while being immersed in virtual reality
No full textImmersive virtual reality (IVR) is an artificially designed environment that can be used to produce realistic and engaging environments which are being used actively in the field of healthcare through training and rehabilitation. The use of IVR nowadays ranges from training surgical operations in a safe environment to neurorehabilitation. IVR implementation for rehabilitation is more task-specific, enhances patients' attention during training, and provides visual feedback. Avatar, a virtual extension of the user, can be used to interact with the virtual environment and aid in postural adjustments during rehabilitation. Although IVR training of upper limbs is often seen, the research is ongoing for lower limbs. Walking activities in the real-world post-stroke are essential for active participation in the community and to reduce potential mental illness. There is ongoing research on how to implement walking activities in VR. This study aimed to explore the effect of visualizing different ground surfaces on gait patterns. Twelve healthy young participants were recruited for the experiment. Two scenes with a different ground surfaces -- ice and concrete -- were designed. A male and a female avatar were animated and implemented in the scene. The participants were asked to walk eight times in both. Trackers located at the left and right foot and pelvis were used to obtain kinematic parameters such as stride and step length and gait speed. The participants were asked to answer an embodiment questionnaire, which consisted of questions about body ownership, sense of agency, and location, after each scene. We found that the first kinematic values of stride and step lengths and gait speed were lower while walking over the virtual ice scene compared to concrete. Overall, the values of body ownership, and sense of agency were higher when compared to the control questions of body ownership and sense of agency, after each scene. The value of the sense of location after each scene was also higher. The present findings show that the participants embodied in both the scenes and the ground surface had a significant influence on their gait modification. Thus, implementing ground surfaces along with IVR in rehabilitation can benefit patients by helping them adapt their gait to the ground surface.Biomedical Engineerin
Application of convolutional neural network for leukocyte quantification from a smartphone based microfluidic biosensor
No full textAdvancements in computer vision methodologies and machine learning in the medical domain have played a major role in diagnostics and clinical pathology. Cell quantification from whole blood can aid in detecting and managing infections, cardiovascular diseases and biomarker detection which in turn helps in understanding the immunological and genetic disorders, cancers, etc. Developing a point-of-care solution for this will accelerate the therapy timeline and increase the accessibility across the world. Our lab has previously developed a smartphone based microfluidic biosensor for capturing the microscopic images of various components of the blood cells. Using this design, in this study, a deep learning-based cell quantification from the captured images is investigated and the cell counts are predicted using a convolutional neural network architecture. The proposed methodology was evaluated on a dataset varying in numbers, clarity, smartphones, fluorophores and cell numbers. This model was then integrated into an Application Programming Interface (API) to predict the cell counts from an image using the trained model. Our results showed successful prediction of cell counts from a smartphone captured image in cross-validation with R2 = 0.99 for N=33. This helps in eliminating the need for manual pre-processing of an image and morphological methods for cell counting which is a user-skill based approach. This proposed Deep Learning based cell quantification has shown agility and more automated process when compared to the benchmark techniques.M.S.Includes bibliographical reference
A Planar Wideband Wide-Scan Phased Array: Connected Array Loaded with Artificial Dielectric Layers
No full textWe present a novel concept for wideband, wide-scan phased array applications. The array is composed by connected-slot elements loaded with artificial dielectric superstrates. The proposed solution consists of a single multi-layer planar printed circuit board (PCB) and does not require the typically employed vertical arrangement of multiple PCBs. This offers advantages in terms of complexity of the assembly and cost of the array. We developed an analytical method for the prediction of the array performance, in terms of active input impedance. This method allows to estimate the relevant parameters of the array with a negligible computational cost. A design example with a bandwidth exceeding one octave (VSWR<2 from 6.5 to 14.3 GHz) and scanning up to 50 degrees for all azimuth planes is presented.Accepted author manuscriptTera-Hertz SensingElectronic
Efficacy of vitamin D gel in curbing and curing radiation-induced oral mucositis
No full textPurpose of the Study: To assess the effectiveness of an oral gel containing Vitamin D in the therapy and prevention of radiation-induced oral mucositis. Materials and Methods: Sixty head and neck cancer patients seeking radiation therapy agreed to participate in a randomized control clinical trial. The first group consisted of conventional treatment augmented with topical application of oral Vitamin D gel. In the second group, only topical oral Vitamin D gel was prescribed. All patients had clinical evaluations for pain and WHO mucositis scores subsequently for two, four, and six weeks after the start of radiation. Results: After six weeks of radiation, patients in both groups experienced complete remissions or less oral mucositis, with the combination group showing better results. Both groups experienced pain relief, with 83.3% of patients in group 1 experiencing complete remission. Conclusion: Topical oral Vitamin D gel reduced the severity of oral mucositis and mitigated pain when implemented in tandem with conventional therapy
Ensemble Learning on Deep Neural Networks for Image Caption Generation
Get PDFabstract: Capturing the information in an image into a natural language sentence is
considered a difficult problem to be solved by computers. Image captioning involves not just detecting objects from images but understanding the interactions between the objects to be translated into relevant captions. So, expertise in the fields of computer vision paired with natural language processing are supposed to be crucial for this purpose. The sequence to sequence modelling strategy of deep neural networks is the traditional approach to generate a sequential list of words which are combined to represent the image. But these models suffer from the problem of high variance by not being able to generalize well on the training data.
The main focus of this thesis is to reduce the variance factor which will help in generating better captions. To achieve this, Ensemble Learning techniques have been explored, which have the reputation of solving the high variance problem that occurs in machine learning algorithms. Three different ensemble techniques namely, k-fold ensemble, bootstrap aggregation ensemble and boosting ensemble have been evaluated in this thesis. For each of these techniques, three output combination approaches have been analyzed. Extensive experiments have been conducted on the Flickr8k dataset which has a collection of 8000 images and 5 different captions for every image. The bleu score performance metric, which is considered to be the standard for evaluating natural language processing (NLP) problems, is used to evaluate the predictions. Based on this metric, the analysis shows that ensemble learning performs significantly better and generates more meaningful captions compared to any of the individual models used.Dissertation/ThesisMasters Thesis Software Engineering 201
Utility of three dimensional (3-D) ultrasound and power Doppler in identification of high risk endometrial cancer at a tertiary care hospital in southern India: A preliminary study
No full textObjective: The study was conducted to find the utility of three dimensional (3-D) ultrasound and Doppler sonography in differentiating benign and malignant endometrial lesions and to ascertain the association of sonology parameters with type, grade and stage of endometrial cancer. Materials and methods: Women attending the gynaecology department of a tertiary care hospital, with a provisional diagnosis of carcinoma endometrium were subjected to three dimensional power Doppler ultrasound evaluation and assessment of vascular patterns. VOCAL (Virtual Organ Computer-aided Analysis) software was used to assess volume, Vascularisation Index (VI), Flow Index (FI) and Vascularisation Flow Index (VFI). Ultrasound parameters were compared with histologic diagnosis to evaluate the diagnostic performance using Receiver Operating Characteristic (ROC) Curve. Results: Sixty-four women were included in the study, 33 with benign and 31 with malignant endometrial lesions. Larger endometrial volume and higher Doppler indices correlated with malignant lesions. The variables with good discriminatory potential between benign and malignant status were VI and VFI, having a sensitivity of 90.3% and specificity of around 80%. VFI (adjusted odds ratio of 40.4; (95% CI – 8.46–192.88), p value < 0.001) was the only significant variable identified by multivariate logistic regression, when adjusted for age and post-menopausal status. Multiple global and focal vessel pattern was seen predominantly in malignant cases (specificity 93.9%), although the sensitivity was low (61.2%). Higher stages and grades of tumour and non-endometrioid types had higher Doppler indices, and requires further evaluation. Conclusions: 3-D ultrasound has good discrimination potential between benign and malignant endometrial lesions and could be useful as a screening tool. However, utility of 3-D tool for differentiation between tumour characteristics needs further validation. Keywords: Endometrial cancer, Three dimensional ultrasound, Power Doppler, Vessel patterns, Vascularisation flow inde
K-Metro domination number of slanting ladder graph
Get PDFA “dominating set D of a graph G = G(V, E) is called Metro dominating set of G. If for every pair of vertices u, v there exists a vertex in D such that d(u,w) ≠ d(v, w). The K-Metro domination number of slanting ladder graph (ꝩβk (S(Ln))), is the order of smallest K-dominating set of S(Ln) which serves as a matric set. In this paper we calculate K-Metro domination number of slanting ladder graph (ꝩβk (S(Ln)))
- …
