Commerce and Management
Permanent URI for this communityhttp://repository.kln.ac.lk/handle/123456789/140
Browse
Item Feature Extraction from Sub-Decimeter Resolution Images Using Convolutional Neural Networks(5th International Conference for Accounting Researchers and Educators (ICARE – 2019), Department of Accountancy, Faculty of Commerce & Management Studies, University of Kelaniya, Sri Lanka, 2019) Aarthy, S.L.; Sujatha, R.Deep Learning has a wide sphere in analysis of the problem. One such problem which is in high demand for research is the extraction of features and pixel-level classification of aerial images which requires the ability to learn the concepts from spatial data. The aim of this paper is to use Convolution Neural Networks to learn those variations by using the state-of-the-art downsample-then-upsample architecture. The overall goal of labeling every pixel of the original resolution is achieved through this architecture. The results show that the overall accuracy is good; there is an improvement in the predicted geometric accuracy and during the inference time the efficiency is also high. The proposed architecture is tested on Potsdam sub-decimetre resolution dataset which is given by the ISPRS and it comprises many annotated tiles for the evaluation of systems using spatial data in an unbiased way.Item Handwritten Character Recognition Using Learning Vector Quantization(5th International Conference for Accounting Researchers and Educators (ICARE – 2019), Department of Accountancy, Faculty of Commerce & Management Studies, University of Kelaniya, Sri Lanka, 2019) Sujatha, R.; Aarthy, S.L.Optical Character recognition is a very futile area of research in the field of image processing. Handwritten character recognition is the most challenging domain of OCR because every person tends to have his/her writing style. As a result, there is variance in every sample input taken from different users. Due to the presence of no standalone handwriting template and huge diversity of people's writing styles, an adaptive and effective character recognition module is required for efficiently identifying handwritten characters. On the other hand, Learning Vector Quantization or LVQ is a kind of supervised neural network which can learn and remember if proper training is provided. This paper focuses on constructing a Learning Vector Quantization based handwritten character recognition module which will be able to effectively identify different handwriting styles and recognize them with a significantly high degree of accuracy.