DHANASEKARAN
@mailamengg.com
Associate Professor and Department of Electronics and Communication Engineering
MAILAM ENGINEERING COLLEGE
RESEARCH INTERESTS
Image Processing,
VLSI,
MEMS
NEMS
Artificial Intelligent,
Robotics
Scopus Publications
Scopus Publications
Gudapati Ashoka Chakravarthy, M. Thiruppathi, S. Kandasamy, and K. Dhanasekaran
ANSF Publications
Crop residue incorporation is a key component of sustainable cropping systems. It reduces the adverse effects of residue burning and enhances soil fertility. Effective usage of crop residue in the field and proper management are required. With this background, a field experiment was conducted during 2020 – 21 in the maize-rice cropping sequence at Annamalai University Experimental Farm, Department of Agronomy, Faculty of Agriculture, Annamalai University, Annamalai Nagar, Tamil Nadu to find out the residual effect of different mulching practices adopted in preceding maize crop and maize stubble incorporation on the growth, yield and economics of transplanted rice Var. ADT 46. The field experiment was conducted in Factorial Randomized Block Design with three replications. In factor I, soil moisture conservation in preceding maize crop viz., mulching of sugarcane trash, water hyacinth, hydrogel and control were allotted. In factor II, different levels of maize crop stubble incorporation on rice viz., 0, 33, 66 and 100% were provided. Water hyacinth mulched plot (M3) to the preceding crop registered significantly (þ <0.05) higher yield parameters, yield and economic returns of succeeding rice. The lower values were observed in unmulched (M1) plot. With respect to maize crop stubble incorporation on rice crop, the incorporation of 66% (SI3) of maize stubble registered higher yield parameters, yield and economic returns. In the interaction effects, mulching with water hyacinth to preceding maize + maize crop stubble incorporation at 66% in rice crop (M3SI3) recorded significantly (þ <0.05) higher yield parameters, yield and economic returns than other treatments. The lowest values were recorded with an unmulched + 100% crop residue incorporated (M1SI4) plot. Mulching the preceding maize crop with water hyacinth at a rate of 12 t ha-1 and incorporating 66% maize stubble into the transplanted rice (M3SI3) had a remarkable yield advantage and financial rewards.
S. Ganesan, K. Dhanasekaran, and N. Nishavithri
IEEE
Because of the outstanding headway of data innovation, PC, capacity frameworks and systems administration innovation, clinical gadgets and clinical conclusion has gained huge prominence over the most recent twenty years. For the most part in clinical field including biomedical science, the impact of such advances is getting evident, permitting the discovery and determination in a significantly more striking manner. The critical obstacle in the indicative imaging study is to choose a picture with no significant subtleties being lost. All things considered, over the span of recovery or again resulting preparing stages, the information caught can be misshaped by commotions or curios. Clamor is characterized as the underlying pixel esteem being changed aimlessly. Commotion brings down the lucidity of the picture which is especially significant at whatever point the constructions are checked is more modest and indeed, even has relatively helpless force. De-noising of picture information is in this manner significant, and in clinical diagnostics it has consistently been an important pre-preparing level. An investigation of a few critical investigates in the field of picture de-noising is examined in this article. Since pictures were very fundamental in any space, picture de-noising is for sure a significant preprocess earlier towards more picture investigation, such as division, extraction of highlights, surface investigation, and so forth This examination proposed to play out the far reaching investigation of different de-noising systems for clinical imaging that includes MRI, CT and Retinal fundus pictures. An examination concentrate with many existing strategies approaches zeroed in on similarity tests, uncovers that the proposed approach is better in picture consistency than them.
Deepak V., M. Rajesh Khanna, K. Dhanasekaran, P. G. Om Prakash, and D. Vijendra Babu
IEEE
In all the technological fields, the data size increases very rapidly and also database becomes very bulk in size. Users using bulk databases confront several challenges, such as determining which query produces the most relevant results. As the number of users has increased dramatically in recent years, there have been various competitions for recommendation systems. For enhancing or building recommendation systems all most or commonly everybody come up with an idea of collaborative filtering technique. When database or the data size increases it also reflect the processing time consumed and as well as the proposals will have potential. It is the best errand to give proposal to huge scope issues to create high greatness suggestions. Nonetheless, several approaches for the expansion of the recommender framework have been presented. Perhaps, the most and famous popular framework in for modern large datasets is Map Reduce, due to the outstanding features as gullibility, fault-tolerance, ease and effective of programming, flexibility. This paper aims to state the enlightening the status of effective and parallel query processing using Apache Mahout, Map Reduce and collaborative filtering.
K. Dhanasekaran, P. Anandan, and N. Kumaratharan
Wiley
P Varalakshmi, Gowtham Yuvaraj, Karthiga Dhanasekaran, and Kalaiyarasi Samannan
IEEE
The evolution of image processing technologies over the recent years along with deep learning has contributed tremendously towards the prediction of many diseases using the training models. Reducing false positive rate and false negative rate to a significant amount results in better outcomes of classification and makes it more reliable to use. Smoking, infections, pollution and genetics prove to be the major causes for lung diseases and they result in affecting interstitium, blood vessels, chest wall, alveoli, pleura and trachea (windpipe). Upon the early symptoms of lung diseases like chronic cough, shortness of breath, wheezing and coughing up blood, the diagnosis becomes inevitable. The medical images of lungs like CT scan, MRI, X-rays etc to detect diseases can be utilized to feed neural networks for training by labelling. Various aspects like better pre-processing methodologies, image reconstruction and image enhancement lead to a good quality of images being given to the neural networks as inputs. As far as the neural network is concerned, choosing appropriate number of hidden layers and effective activation function are expected to yield finer solution in our motto of reducing miss-rate and fall-out. A multitude of neural networks with different activation function is already being proposed paving way for more research on the area to produce further enhancement in the field. Over the recent years, deep learning has definitely emerged as the state of the art, contributing to an improved performance in a variety of medical applications. In this paper, we have proposed few CNN architectures that differ from one another by hyper parameters to classify the images based on Pneumonia affected or not.
K. Dhanasekaran and R. Priyarani
Springer Netherlands
A field experiment was conducted to study the effect of various sources of boron on the growth and yield of cotton. The experiment was laid out in factorial randomized block design with 13 treatments. The treatments consisted of three levels of boron supplied through four sources with one control. Three levels of boron were (L1) 2.5, (L2) 5.0, and (L3) 7.5kgha−1. The required quantity of boron for the treatments was supplied through four sources,viz., (S1) borax, (S2) boric acid, (S3) sodium tetraborate, and (S4) calcium boro-humate. The experiment was conducted in a Typic Ustifluvent soil having the pH 7.8, EC 1.30 dSm−1, and CEC 11.70 c mol [P+] kg−1. The soil was low in organic carbon (4.5gkg−1) and available N (198.14kgha−1), medium in available P (12.03kgha−1), high in available K (310kgha−1), and deficient in available (hot water soluble) B (0.47mgkg−1). Cotton var. MCU 7 was grown as test crop. The result of the experiment clearly revealed that application of B through four different sources showed a significant increase in growth and yield of cotton over control. Among the four sources, addition of B through calcium boro-humate showed a superior performance in respect of growth and yield as compared to the other three sources. Based on their efficiency, the four sources of boron tested in the experiment are given in the descending order as calcium boro-humate>sodium tetrahumate>boric acid>borax. Application of boron at 7.5kgha−1through calcium boro-humate increased the seed cotton yield to the tune of 21.20%, besides improving availability of boron in postharvest soil.