CNN ensemble approach for early detection of sugarcane diseases – a comparison K J Kavitha, K Krishna Prasad International Journal of Electronics and Telecommunications, 2024 This paper mainly concentrates and discusses on sugarcane crop, the variety of cane seeds available for sowing; various cane diseases and its early detection using different approaches. Machine Learning (ML) and Deep Learning (DL) techniques are used to analyze agricultural data like temperature, soil quality, yield prediction, selling price forecasts, etc. and avoid crop damage from a variety of sources, including diseases. In the proposed work, with particular reference to eight specific sugarcane crop diseases and including healthy crop database, the neural network algorithms are tested and verified in terms quality metrics like accuracy, F1 score, recall and precision.
Compressed Sensing Reconstruction Algorithms for Medical Images – A Comparison K. J. Kavitha, Vishwaraj B. Manur, P. G. Suprith, Mahendra S. Naik, S. N. Chaitra Smart Hospitals 5g 6g and Moving Beyond Connectivity, 2024 The Internet of Things (IoT) is a rapidly developing field of technology that has the potential to revolutionize healthcare, smart cities, daily human activities, manufacturing and other industries. Incorporating compressive sensing (CS) into the design of IoT platforms is a highly appealing paradigm. The medical imaging community has been very interested in CS because of its potential to produce high-quality picture reconstructions with less data by taking use of compressibility. In this paper, we have discussed various compressing reconstruction algorithms that could be used in the medical imaging application for better image construction by eliminating unnecessary redundant bits. The orthogonal matching pursuit algorithm (OMP), sparsity-adaptive matching pursuit algorithm (SaMP), adaptive step-size SaMP algorithm (AS-SaMP) and dynamic-step-size SaMP (DSS-SaMP) are discussed, evaluated and compared with each other in terms of bit error rate (BER), signal error rate (SER), and mean square error (MSE).
Patient Monitoring Using 5G, with MIMO-NOMA for mm-Wave Communications in Heterogeneous Networks P. G. Suprith, Mohammed Riyaz Ahmed, Mahendra Shridhar Naik, K. J. Kavitha, S. N. Chaitra Smart Hospitals 5g 6g and Moving Beyond Connectivity, 2024 Future Internet of Things (IoT) applications are expected to leverage modular computing at the edge (MCE), an approach that is thought to be favorable, to support computation-intensive and time-sensitive tasks. Wireless network areas are defined by their precision in transmitting and receiving data, and they operate either inside or close to the human body. It needs to use less energy and operate for extended periods of time. The short broadcast and receiving distance of less than a few meters is one of its issues. In order to address this issue, researchers focused on a simulation based on a fifth generation (5G) network using multiple-input/multiple-output (MIMO) terminals that enable transport signals to the center (medical center) at 2.4, 2.8, or 5 GHz bands. The power is allocated to the patients through 5G non-orthogonal multiple access (NOMA) topology network. Using Bit Error Rate (BER), achievable total rate, outage likelihood, and the efficiency of the patient's power allocation are assessed. We have seen that devices considerably prolong the life of the nodes, operate better, and consume less energy based on the data acquired, while the reaction time for data transfers was accelerated by 5G devices. Furthermore, the advantage of MIMO antennas is that they increase the long-term reliability of the connection among locations.
An efficient medical image watermarking technique using integer wavelet transform and quick/fast response codes K.J. Kavitha, Priestly B. Shan International Journal of Intelligent Systems Technologies and Applications, 2019 Securing the medical images to make it tamper free is a terribly difficult task. With the advanced digital watermarking (DWM) technique, we are able to protect the medical images by evaluating validation, dependability, privacy and integrity. The DWM is implemented in two main domains: transform and spatial. The DWM is mostly implemented using the transform techniques such as singular valued decomposition (SVD), discrete cosine transform (DCT), discrete wavelet transform (DWT), integer wavelet transform (IWT) and combination of these techniques. One of the foremost challenges in these technologies is information embedding capability and this parameter is considered for evaluation of the system. One of the possible ways to reduce the number of embedding bits in information is to use quick response code (QR). The proposed DWM system uses: IWT, bit plane and QR code.
