Shankar M. Krishnan

Region labeling of colonoscopic images using fuzzy logic

A fuzzy rule base approach to the labeling of colonoscopic images to render assistance to the clinician is presented. The color images are segmented using a scale-space filter. Several features are selected and fuzzified. The knowledge-based fuzzy rule base system performs the labeling of the segmented regions into background, lumen, and abnormalities (polyps, bleeding lesions). The system has been tested on several precaptured colonoscopic images. The results are encouraging. Modifications are in progress to prepare for real-time clinical testing.

Collaboration for cooperative work experience programs in biomedical engineering education

Incorporating cooperative education modules as a segment of the undergraduate educational program is aimed to assist students in gaining real-life experience in the field of their choice. The cooperative work modules facilitate the students in exploring different realistic aspects of work processes in the field. The track records for cooperative learning modules are very positive. However, it is indeed a challenge for the faculty developing Biomedical Engineering (BME) curriculum to include cooperative work experience or internship requirements coupled with a heavy course load through the entire program. The objective of the present work is to develop a scheme for collaborative co-op work experience for the undergraduate training in the fast-growing BME programs. A few co-op/internship models are developed for the students pursuing undergraduate BME degree. The salient features of one co-op model are described. The results obtained support the proposed scheme. In conclusion, the cooperative work experience will be an invaluable segment in biomedical engineering education and an appropriate model has to be selected to blend with the overall training program.

Academic program models for undergraduate biomedical engineering.

There is a proliferation of medical devices across the globe for the diagnosis and therapy of diseases. Biomedical engineering (BME) plays a significant role in healthcare and advancing medical technologies thus creating a substantial demand for biomedical engineers at undergraduate and graduate levels. There has been a surge in undergraduate programs due to increasing demands from the biomedical industries to cover many of their segments from bench to bedside. With the requirement of multidisciplinary training within allottable duration, it is indeed a challenge to design a comprehensive standardized undergraduate BME program to suit the needs of educators across the globe. This papers objective is to describe three major models of undergraduate BME programs and their curricular requirements, with relevant recommendations to be applicable in institutions of higher education located in varied resource settings. Model 1 is based on programs to be offered in large research-intensive universities with multiple focus areas. The focus areas depend on the institutions research expertise and training mission. Model 2 has basic segments similar to those of Model 1, but the focus areas are limited due to resource constraints. In this model, co-op/internship in hospitals or medical companies is included which prepares the graduates for the work place. In Model 3, students are trained to earn an Associate Degree in the initial two years and they are trained for two more years to be BMEs or BME Technologists. This model is well suited for the resource-poor countries. All three models must be designed to meet applicable accreditation requirements. The challenges in designing undergraduate BME programs include manpower, facility and funding resource requirements and time constraints. Each academic institution has to carefully analyze its short term and long term requirements. In conclusion, three models for BME programs are described based on large universities, colleges, and community colleges. Model 1 is suitable for research-intensive universities. Models 2 and 3 can be successfully implemented in higher education institutions with low and limited resources with appropriate guidance and support from international organizations. The models will continually evolve mainly to meet the industry needs.

Application of Analytics to Big Data in Healthcare

In the current age of smart phones and wearable devices, vast amounts of patient health data files forming Big Data are being placed into large databases where they can be accessed by multiple users including doctors, caregivers and patients. The estimated spending on healthcare in 2015 in the U.S. is around

Kernel Machines for Imbalanced Data Problem in Biomedical Applications

3.2 trillion, which triggers the question of improvement of patient care while containing the costs. The objective of the present study is to review a few applications of analytics of Big Data in the healthcare field and the associated outcomes. Big Data is generally characterized by the volume, velocity, variety and veracity of complex data. Many hospitals have applied analytics to big data from various sources including patient health records to achieve overall improvement in healthcare. Operationally, most of the pertinent data of patients are made available on demand so doctors can see how other treatments have worked globally and apply relevant results to facilitate better decision making and interventions. Making proper use of big data analytics in healthcare can lead to improvement in care delivery coupled with significant cost savings. Concurrent challenges to be addressed include accessibility, privacy, security, usability, implementation costs, transportability, interoperability, and standardization. In conclusion, employing efficient and streamlined analytics to big data will contribute to quick and accurate diagnosis, appropriate treatment, reduced costs and improved overall healthcare quality.

Project-based learning with international collaboration for training biomedical engineers

Kernel machines such as the support vector machines (SVMs) have been reported to perform well in many applications. However, the performance of a binary SVM can be adversely affected by an imbalanced set of training samples, known as the imbalanced data problem. One-class SVMs, as a recognition-based approach, can be used to train and recognize the majority class and such kernel machines have already been developed. In this chapter, we review and study the effects of imbalanced datasets on the performance of both one-class SVMs and binary SVMs. We show that a hybrid kernel machine comprising one-class SVMs and binary SVMs in a multi-classifier system alleviates the imbalanced data problem. We also report the deployment of such hybrid kernel machines in two biomedical applications where the imbalanced data problem exists.

Assessment of Distance Measurement with Selected Wearable Devices in Telemonitoring

Training biomedical engineers while effectively keeping up with the fast paced scientific breakthroughs and the growth in technical innovations poses arduous challenges for educators. Traditional pedagogical methods are employed for coping with the increasing demands in biomedical engineering (BME) training and continuous improvements have been attempted with some success. Project-based learning (PBL) is an academic effort that challenges students by making them carry out interdisciplinary projects aimed at accomplishing a wide range of student learning outcomes. PBL has been shown to be effective in the medical field and has been adopted by other fields including engineering. The impact of globalization in healthcare appears to be steadily increasing which necessitates the inclusion of awareness of relevant international activities in the curriculum. Numerous difficulties are encountered when the formation of a collaborative team is tried, and additional difficulties occur as the collaboration team is extended to international partners. Understanding and agreement of responsibilities becomes somewhat complex and hence the collaborative project has to be planned and executed with clear understanding by all partners and participants. A model for training BME students by adopting PBL with international collaboration is proposed. The results of previous BME project work with international collaboration fit partially into the model. There were many logistic issues and constraints; however, the collaborative projects themselves greatly enhanced the student learning outcomes. This PBL type of learning experience tends to promote long term retention of multidisciplinary material and foster high-order cognitive activities such as analysis, synthesis and evaluation. In addition to introducing the students to experiences encountered in the real-life workforce, the proposed approach enhances developing professional contracts and global networking. In conclusion, despite initial challenges, adopting project-based learning with international collaboration has strong potentials to be valuable in the training of biomedical engineering students.

Embedding Internship Programs to Augment BME Education

The use of wearable devices in health applications is not only being utilized in illness, it is also a major area of interest in fitness. Wearable devices for fitness tracking are available to consumers and can measure and calculate important fitness data trends based on movement and physiological parameters. Accelerometers are the sensors that provide information on the movement parameters of step count and distance. It is understood to obtain a distance measurement, there must be information provided about the step-to-step length but these devices typically only measure step count, and do provide the step-to-step length. The step-to-step length is based on the users height which is input into the settings, allowing for the calculation of the distance. The intention of this study is to introduce a preliminary experiment to evaluate two different devices -- the Fitbit Flex and the Polar Loop -- for distance measurement methods, based on the step-to-step length values determined from collected device data. One measurement method uses a default stride length while the other method uses a stride speed algorithm. The step-to-step values for each device were compared to an experimental ground truth value for accuracy, which was used as a pilot for comparing the initial device data. In this preliminary study, the default stride length method provided a more accurate method for measuring the distance. A subsequent study is expected to be performed with more participants and parameters, which will be more suitable for providing a better comparison between the two devices against a ground truth.

The need of ethics training for biomedical engineering students

For undergraduate programs in biomedical engineering (BME), a comprehensive base of multiple disciplines required for BME studies on campus has been a big challenge, the Industry Professional Advisory Committee members usually recommend incorporating experiential learning modules of co-op or internship in the curriculum. Embedding cooperative modules within the undergraduate BME educational program is aimed to assist students in gaining the highly valuable real-life experience. The internship modules facilitate the students in exploring different realistic aspects of the complex work processes in the biomedical engineering field. It must be emphasized that different countries follow different models for BME education. At the international level, the developers of the BME curriculum find the inclusion of internship experience or internship with a heavy course load in the program a formidable challenge. Having a single model is not likely to work in different countries. The main objective of the present work is to develop cooperative experiential learning models for BME undergraduate students that can be applied internationally and to propose multiple partner organizations to host the co-op. In this paper, designs of a few co-op/internship models embedded in the undergraduate BME curriculum and an innovative array of co-op hosting organizations are described. The results obtained clearly support the proposed co-op/internship scheme. In conclusion, integrating the internship experience will be of significant value in biomedical engineering education by giving opportunities for real-life work experience to the students. For sustained success at the international level, it is essential that a suitable model must be selected to blend with the mission of the overall training program at the academic institution.

Exoskeleton systems kinematics analysis with graph-matroid approach

Biomedical sciences is a rapidly growing field and due to its interdisciplinary nature, it presents many unique ethical challenges. Students in biomedical sciences and engineering need training in bioethics so that they are adequately prepared to face many ethical challenges that they will face in their future career as they develop new drugs and devices which may transform the lives of our future patients.