Deep Learning Based Compression with Classification Model on CMOS Image Sensors

Palani, Vinayagam; Thanarajan, Tamilvizhi; Krishnamurthy, Anand; Rajendran, Surendran

doi:10.18280/ts.400332

Cited by 4 publications

(2 citation statements)

References 20 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Essential measurement of vital signs and physiological functions will be useful. The Android application displayed in the doctor-related system allows him to receive the patient's current status without physicalizing the smartphone's current demographic data [17,18].…”

Section: Related Workmentioning

confidence: 99%

Sensors Based Optimized Closed Loop Control Algorithm to Minimize Hypoglycemia/Hyperglycemia using 4-Variate Time Series Data

G.,

Sundarapandi,

et al. 2024

JNMES

View full text Add to dashboard Cite

The goal of this work is to create a closed-loop control system that combines continuous glucose, carbohydrates, and physiological variable readings to regulate glucose levels to treat hyperglycemia and prevent hypoglycemia, as well as a hypoglycemia early alarm module. We use an Extended Kalman filter (EKF) to estimate time-series coefficients of type 1 diabetic patient's glucose levels using a 4-variate time series data such as glucose level, insulin dose, physical activities, and food consumption. An adaptive Kalman Filter algorithm is best suitable for sensor fusion and also for real-time data, which uses the series of measurement data over the period to tend to predict the unknown variable. We proposed to provide the mealtime bolus as a prolonged bolus, with a slice of the insulin dosage delivered before food and the remainder delivered after food. Our research entails the use of a glucose monitor (CGM), physical activity monitor (accelerometer sensor), carbohydrates monitor (biosensor), an automatic insulin infusion controller that calculates the quantity of insulin to be injected, and an insulin infusion drive. Low-Power and Lossy Routing (LPLR) with 6LoWPAN are proposed for efficient routing and private networks. Our model is evaluated using the UVa/Padova simulator for 25 patients. We conducted our experiments with 25 virtual patients with a mix of all age categories. The simulator's default sample time is one minute, but we have set the 5-minute sampling time. Overall, the proposed models are good at predicting hypoglycaemic (<70 mg/dl), normal glycaemic (>70 and 180 mg/dl), and hyperglycaemic (180 mg/dl) blood sugar levels.

show abstract

Section: Related Workmentioning

confidence: 99%

Sensors Based Optimized Closed Loop Control Algorithm to Minimize Hypoglycemia/Hyperglycemia using 4-Variate Time Series Data

G.,

Sundarapandi,

et al. 2024

JNMES

View full text Add to dashboard Cite

show abstract

“…Because the packet is temporarily kept in the gateway, it is simple to resend it. When the server admits the positive transmission, the packet is destroyed [33,34].…”

Section: Encryptionmentioning

confidence: 99%

SFoG-RPI: A Secured QoS Aware and Load Balancing Framework for FoG Computing in Healthcare Paradigm

Gunasekar,

Krishnamurthy,

Thanarajan

et al. 2023

RIA

View full text Add to dashboard Cite

Diabetes, characterized by persistently high blood glucose levels, has been identified as a hazardous health condition, potentially leading to severe complications such as heart attacks, strokes, and heart failure. This study introduces a fog-based remote health monitoring system designed to mitigate the devastating impacts of diabetes and hypoglycemia. This system persistently monitors health parameters including glucose levels, carbohydrate intake, physical activities, heart rate, and blood pressure. It additionally supports advanced services such as feature extraction, distributed local storage, and enhanced security. The traditional cloud-based architecture, while effective, often results in significant latency due to the processing of vast amounts of data. By bringing computing servers closer to users, Fog computing addresses this issue, reducing latency, and increasing security, resource accessibility, and on-demand scaling. In this context, the proposed system aims to minimize latency and network usage while addressing critical issues such as security, access control, and privacy. It employs lossy data compression at the gateway level to decrease network bandwidth and enhance efficiency. Furthermore, the system introduces a novel Load Balancing mechanism to distribute the load among fog nodes evenly. It utilizes lightweight cryptographic algorithms, efficient key exchange protocols, and digital signatures to ensure confidentiality, authentication, and user privacy. The performance of the proposed framework was evaluated in terms of average processing time, energy consumption management, computational resource distribution, latency, and network usage. When compared with other systems, the proposed framework demonstrated superior results, thus validating its effectiveness.

show abstract