scispace - formally typeset
Search or ask a question
Author

Maheshi B. Dissanayake

Other affiliations: University of Surrey
Bio: Maheshi B. Dissanayake is an academic researcher from University of Peradeniya. The author has contributed to research in topics: Network packet & Codec. The author has an hindex of 7, co-authored 36 publications receiving 157 citations. Previous affiliations of Maheshi B. Dissanayake include University of Surrey.

Papers
More filters
Journal ArticleDOI
TL;DR: In this paper, a threefold deep learning architecture is proposed for tumor extraction and segmentation of tumor boundaries correctly, which includes a deep convolutional neural network (CNN), a region-based CNN and a Chan-Vese segmentation algorithm.
Abstract: One of the main requirements of tumor extraction is the annotation and segmentation of tumor boundaries correctly. For this purpose, we present a threefold deep learning architecture. First, classifiers are implemented with a deep convolutional neural network (CNN) and second a region-based convolutional neural network (R-CNN) is performed on the classified images to localize the tumor regions of interest. As the third and final stage, the concentrated tumor boundary is contoured for the segmentation process by using the Chan-Vese segmentation algorithm. As the typical edge detection algorithms based on gradients of pixel intensity tend to fail in the medical image segmentation process, an active contour algorithm defined with the level set function is proposed. Specifically, the Chan-Vese algorithm was applied to detect the tumor boundaries for the segmentation process. To evaluate the performance of the overall system, Dice Score, Rand Index (RI), Variation of Information (VOI), Global Consistency Error (GCE), Boundary Displacement Error (BDE), Mean Absolute Error (MAE), and Peak Signal to Noise Ratio (PSNR) were calculated by comparing the segmented boundary area which is the final output of the proposed, against the demarcations of the subject specialists which is the gold standard. Overall performance of the proposed architecture for both glioma and meningioma segmentation is with an average Dice Score of 0.92 (also, with RI of 0.9936, VOI of 0.0301, GCE of 0.004, BDE of 2.099, PSNR of 77.076, and MAE of 52.946), pointing to the high reliability of the proposed architecture.

43 citations

Proceedings ArticleDOI
01 Mar 2019
TL;DR: A Convolutional Neural Network (CNN), for classification problem and Faster Region based convolutional neural network (Faster R-CNN) for segmentation problem with reduced number of computations with a higher accuracy level is proposed.
Abstract: In this study the problem of fully automated brain tumor classification and segmentation, in Magnetic resonance imaging (MRI) containing both Glioma and Meningioma types of brain tumors are considered. This paper proposes a Convolutional Neural Network (CNN), for classification problem and Faster Region based Convolutional Neural Network (Faster R-CNN) for segmentation problem with reduced number of computations with a higher accuracy level. This research has used 218 images as training set and the systems shows an accuracy of 100% in Meningioma and 87.5% in Glioma classifications and an average confidence level of 94.6% in segmentation of Meningioma tumors. The segmented tumor regions are validated through ground truth analysis and manual analysis by a Neurologist.

41 citations

Journal ArticleDOI
TL;DR: This work derives the analytical expression for the approximate bit error probability (BEP) of the diffusion-based MC system with the full absorption receiver and develops the particle-based simulation framework to simulate the proposed system with RS code to verify the accuracy of the derived analytical results.
Abstract: Molecular communication (MC) has recently emerged as a novel paradigm for nano-scale communication utilizing molecules as information carriers. In diffusion-based molecular communication, the system performance is constrained by the inter-symbol-interference caused by the crossover of information carrying molecules in consecutive bits. To cope with this, we propose the Reed-Solomon (RS) codes as an error recovery tool, to improve the transmission reliability in diffusion-based MC systems. To quantify the performance improvement due to RS codes, we derive the analytical expression for the approximate bit error probability (BEP) of the diffusion-based MC system with the full absorption receiver. We further develop the particle-based simulation framework to simulate the proposed system with RS code to verify the accuracy of our derived analytical results. Our results show that, as the number of molecules per bit increases, the BEP of the system with RS codes exhibits a substantial improvement than that of non-coded systems. Furthermore, the BEP of the proposed system with RS codes can be greatly improved by increasing the minimum distance of the codeword.

25 citations

Journal ArticleDOI
TL;DR: This paper presents the first tractable analytical framework for the collective signal strength at a partially absorbing receiver due to the desired transmitter under the impact of a swarm of interfering transmitters in a 3D large-scale MC system using stochastic geometry and proposes Reed–Solomon error correction coding and two types of information molecule modulating scheme.
Abstract: In recent years, communicating information using molecules via diffusion has attracted significant interest in bio-medical applications. To date, most of the studies have concentrated on point-to-point molecular communication (MC), whereas in a realistic environment, multiple MC transmitters are likely to transmit molecular messages simultaneously sharing the same propagation medium, resulting in significant performance variation of the MC system. In this type of large-scale MC system, the collective signal strength at the desired receiver can be impaired by the interference caused by other MC transmitters, which may degrade the system reliability and efficiency. This paper presents the first tractable analytical framework for the collective signal strength at a partially absorbing receiver due to the desired transmitter under the impact of a swarm of interfering transmitters in a 3D large-scale MC system using stochastic geometry. To combat the multi-user interference and the intersymbol interference (ISI) in the multi-user environment, we propose Reed–Solomon (RS) error correction coding, due to its high effectiveness in combating burst and random errors, as well as the two types of information molecule modulating scheme, where the transmitted bits are encoded using two types of information molecules at consecutive bit intervals. We derive analytical expressions for the bit error probability (BEP) of the large-scale MC system with the proposed two schemes to show their effectiveness. The results obtained using Monte Carlo simulations, match exactly with the analytical results, justifying the accuracy of the derivations. Results reveal that both schemes improve the BEP by a factor of 3–4 compared with that of a conventional MC system without using any ISI mitigation techniques. Due to the implementation simplicity, the two-type molecule encoding scheme is better than the RS error correction coding scheme, as the RS error correction coding scheme involves additional encoding and decoding process at both the transmitter and receiver nodes. Furthermore, the proposed analytical framework can be generalized to the analysis of other types of receiver designs and performance characterization in multi-user large-scale MC systems. Also, the two types of information molecule modulating scheme can be extended to M-type of information molecule modulating scheme without loss of generality.

21 citations

Proceedings ArticleDOI
26 Aug 2008
TL;DR: The simulation results under packet loss environments show that the proposed algorithm outperforms the existing redundant picture coding of JSVM, which is based on providing motion vectors as redundant data.
Abstract: This proposal presents a new error robust strategy for encoding redundant pictures for the H.264/AVC standard. The method is based on providing motion vectors as redundant data, i.e. providing extra protection to the motion information of the encoded stream. The proposed system is implemented based on the existing redundant coding algorithm of the scalable extension of H.264/AVC. The performance of the algorithm is evaluated using various objective quality measurements under both error free and error prone Internet protocol (IP) packet network environments. The proposed algorithm increases the bandwidth utilization with slight degradation in the primary picture quality for error free conditions, compared to the existing redundant coding method of JSVM (joint scalable video model). Furthermore, the simulation results under packet loss environments show that the proposed algorithm outperforms the existing redundant picture coding of JSVM.

15 citations


Cited by
More filters
Journal ArticleDOI
TL;DR: Computer and Robot Vision Vol.
Abstract: Computer and Robot Vision Vol. 1, by R.M. Haralick and Linda G. Shapiro, Addison-Wesley, 1992, ISBN 0-201-10887-1.

1,426 citations

Journal ArticleDOI
TL;DR: This survey presents a comprehensive analysis of the exploitation of network slicing in IoT realisation and discusses the role of other emerging technologies and concepts, such as blockchain and Artificial Intelligence/Machine Learning (AI/ML) in network slicing and IoT integration.
Abstract: Internet of Things (IoT) is an emerging technology that makes people’s lives smart by conquering a plethora of diverse application and service areas. In near future, the fifth-generation (5G) wireless networks provide the connectivity for this IoT ecosystem. It has been carefully designed to facilitate the exponential growth in the IoT field. Network slicing is one of the key technologies in the 5G architecture that has the ability to divide the physical network into multiple logical networks (i.e., slices) with different network characteristics. Therefore, network slicing is also a key enabler of realisation of IoT in 5G. Network slicing can satisfy the various networking demands by heterogeneous IoT applications via dedicated slices. In this survey, we present a comprehensive analysis of the exploitation of network slicing in IoT realisation. We discuss network slicing utilisation in different IoT application scenarios, along with the technical challenges that can be solved via network slicing. Furthermore, integration challenges and open research problems related to the network slicing in the IoT realisation are also discussed in this paper. Finally, we discuss the role of other emerging technologies and concepts, such as blockchain and Artificial Intelligence/Machine Learning (AI/ML) in network slicing and IoT integration.

173 citations

Journal ArticleDOI
TL;DR: This paper uses magnetic resonance imaging images to train a new hybrid paradigm which consists of a neural autoregressive distribution estimation (NADE) and a convolutional neural network (CNN) and test this model with 3064 T1-weighted contrast-enhanced images with three types of brain tumors.

96 citations

Patent
Simon Winder1
03 Aug 2012
TL;DR: Point Cloud Smoother as mentioned in this paper provides various techniques for refining a 3D point cloud or other 3D input model to generate a smoothed and denoised 3D output model.
Abstract: A “Point Cloud Smoother” provides various techniques for refining a 3D point cloud or other 3D input model to generate a smoothed and denoised 3D output model. Smoothing and denoising is achieved, in part, by robustly fitting planes to a neighborhood of points around each point of the input model and using those planes to estimate new points and corresponding normals of the 3D output model. These techniques are useful for a number of purposes, including, but not limited to, free viewpoint video (FVV), which, when combined with the smoothing techniques enabled by the Point Cloud Smoother, allows 3D data of videos or images to be denoised and then rendered and viewed from any desired viewpoint that is supported by the input data.

54 citations

Journal ArticleDOI
TL;DR: In this survey, the enabling technologies have been presented to apprehend the state-of-art with the discussion on the possibility of the hybrid technologies and the inter-connectivity of electromagnetic and molecular body-centric nanonetworks is discussed.
Abstract: With the huge advancement of nanotechnology over the past years, the devices are shrinking into micro-scale, even nano-scale. Additionally, the Internet of nano-things (IoNTs) are generally regarded as the ultimate formation of the current sensor networks and the development of nanonetworks would be of great help to its fulfilment, which would be ubiquitous with numerous applications in all domains of life. However, the communication between the devices in such nanonetworks is still an open problem. Body-centric nanonetworks are believed to play an essential role in the practical application of IoNTs. BCNNs are also considered as domain specific like wireless sensor networks and always deployed on purpose to support a particular application. In these networks, electromagnetic and molecular communications are widely considered as two main promising paradigms and both follow their own development process. In this survey, the recent developments of these two paradigms are first illustrated in the aspects of applications, network structures, modulation techniques, coding techniques and security to then investigate the potential of hybrid communication paradigms. Meanwhile, the enabling technologies have been presented to apprehend the state-of-art with the discussion on the possibility of the hybrid technologies. Additionally, the inter-connectivity of electromagnetic and molecular body-centric nanonetworks is discussed. Afterwards, the related security issues of the proposed networks are discussed. Finally, the challenges and open research directions are presented.

50 citations