| dc.description.abstract | Telemedicine, or remote healthcare, is a medical service that enables physicians and
patients to interact and exchange information without being physically present in the same
location. It utilizes digital devices such as computers and smartphones to deliver healthcare
services. Typically, telemedicine involves live video consultations, although some providers
may also offer services through email or text messages. This approach encompasses various
aspects of medical care, including telemedicine consultations, prescriptions, remote
monitoring of health indicators, and remote diagnosis and treatment. Many individuals opt to utilize telemedicine services through their regular healthcare
providers, while others access care through dedicated apps or platforms. Remote consultations
involve the use of technology to exchange information, receive diagnosis, and obtain advice
from healthcare professionals, even when the patient and doctor are physically separated. Telemedicine offers several advantages. It allows patients to save time by avoiding the
need for travel and eliminates the hassles associated with traffic and scheduling conflicts. Additionally, telemedicine reduces the risk of exposure to infectious diseases that may be
present in healthcare settings. However, it is important to note that not all medical cases are
suitable for telehealth consultations. This approach is particularly beneficial for patients who
are located far away from healthcare facilities, elderly or physically weak individuals who
have difficulty walking, individuals who require post-surgical follow-ups, or those with
chronic diseases. In this thesis, The application carried out several key tasks related to telemedicine and
utilizing artificial intelligence models for data collection and disease diagnosis. The
application collected data from patients through remote healthcare IoT devices such as SpO2
(blood oxygen level) measurements, heart rate, and temperature.For each measure we used
compatible devices and technologies to gather these indicators from patients during the
telemedicine process. After collecting the data, the app processed and analyzed the data to
extract useful and relevant information, as well as display it on observational charts. Before integrating an AI algorithm into my application, I conducted thorough testing
and evaluation of various artificial intelligence models for disease diagnosis using datasets
collected from telemedicine in previous projects. These models were developed using
machine learning or deep learning algorithms. The primary goal of this evaluation was to
ensure the accuracy and reliability of the models in remote diagnosis.
7
During the evaluation process, I considered several metrics to assess the usability and
performance of the models. These metrics included accuracy, sensitivity, specificity, and
other relevant measures. By analyzing these metrics, I could determine the effectiveness of
the models in correctly diagnosing diseases based on the collected data. The training and evaluation phase played a crucial role in fine-tuning the AI models
and optimizing their performance. It allowed me to identify any potential limitations or areas
for improvement. This rigorous evaluation process ensured that the selected AI model used in
my application was well-suited for accurate and reliable remote disease diagnosis. In summary, my application aims to harness the advantages of telemedicine, including
convenience and efficiency in delivering healthcare services remotely. It offers numerous
benefits, such as time savings, reduced exposure to infectious diseases, and improved
accessibility for specific patient groups. However, it is crucial to consider the suitability of
telehealth consultations based on individual circumstances and specific medical conditions. By leveraging the capabilities of artificial intelligence and telemedicine, we can enhance
healthcare delivery and ultimately improve patient outcomes. The evaluation of AI models
using telemedicine datasets has provided valuable insights and instilled confidence in the
performance and usability of the chosen model for my application | en_US |
