Research in the recognition of human activities of daily living has significantly improved using deep learning techniques. Traditional human activity recognition techniques often use handcrafted features from heuristic processes from single sensing modality. The development of deep learning techniques has addressed most of these problems by the automatic feature extraction from multimodal sensing devices to recognise activities accurately. In this paper, we propose a deep learning multi-channel architecture using a combination of convolutional neural network (CNN) and Bidirectional long short-term memory (BLSTM). The advantage of this model is that the CNN layers perform direct mapping and abstract representation of raw sensor inputs for feature extraction at different resolutions. The BLSTM layer takes full advantage of the forward and backward sequences to improve the extracted features for activity recognition significantly. We evaluate the proposed model on two publicly available datasets. The experimental results show that the proposed model performed considerably better than our baseline models and other models using the same datasets. It also demonstrates the suitability of the proposed model on multimodal sensing devices for enhanced human activity recognition.

Owing to recent technological advancement, computers and other devices running several image editing applications can be further exploited for digital image processing operations. This paper evaluates various image processing techniques using matrix laboratory (MATLAB-based analytics). Compared to the conventional techniques, MATLAB gives several advantages for image processing. MATLAB-based technique provides easy debugging with extensive data analysis and visualization, easy implementation and algorithmic-testing without recompilation. Besides, MATLAB's computational codes can be enhanced and exploited to process and create simulations of both still and video images. Moreover, MATLAB codes are much concise compared to C++, thus making it easier for perusing and troubleshooting. MATLAB can handle errors prior to execution by proposing various ways to make the codes faster. The proposed technique enables advanced image processing operations such as image cropping/resizing, image denoising, blur removal, and image sharpening. The study aims at providing readers with the most recent MATLAB-based image processing application-tools. We also provide an empirical-based method using two-dimensional discrete cosine transform (2D-DCT) derived from its coefficients. Using the most recent algorithms running on MATLAB toolbox, we performed simulations to evaluate the performance of our proposed technique. The results largely present MATLAB as a veritable approach for image processing operations.

This paper presents a two-way filtering power divider (FPD) with an equal output power ratio of 1:1. This implies that each of the FPD output port would receive 50% of the power at the input port. To achieve miniaturisation, a common square open-loop resonator is used to distribute energy between the two integrated Chebyshev bandpass filters. In addition to distributing energy, the common resonator also contributes one pole to each integrated bandpass filter (BPF), hence, reducing the number of individual resonating elements used in achieving the integrated FPD. To demonstrate the proposed design technique, a prototype FPD centred at 2.6 GHz with a 3 dB fractional bandwidth of 3% is designed, simulated and presented. The circuit model and microstrip layout results of the FPD show good agreement. The microstrip layout simulation responses show that a less than 1.1dB insertion loss and a greater than 16.5dB in-band return loss were achieved. The overall footprint of the integrated FPD is 37mm by 13mm (i.e. 0.32λg x 0.11λg, for λg = guided-wavelength of the 50Ω microstrip line at 2.6 GHz). The integrated FPD reported in this paper shows some promising merits when compared to similar devices recently reported in literature.

No abstract

Adoption techniques are widely applied in and for cloud service usage to improve the slow acceptance rate of cloud services by SMEs. In such context, a well-understood problem is finding a suitable service from the vast number of services offering similar packages to satisfy user requirements such as security, cost, trust and operating systems compatibility has become a big challenge. However, a major drawback of existing techniques such as frameworks, web search, decision support tools, management models, ontology models and agent technology is that they are restricted to a specific task or they replicate service provider offerings. In this paper, we present Cloudysme a cloud service adoption solution, a middleware that is capable of aiding the decision making process for SMEs adoption of cloud services. Using a case study of SaaS storage services offerings by cloud providers, we introduce a new formalism for judging the superiority of one service attribute over another, we propose an extended version of pairwise comparison and Analytical hierarchical Process (AHP) which is a traditional multi-criteria decision method (MCDM) in solving complex comparisons. We solve the issue of service recommendation by introducing an acceptable standard for each service attribute and propose a protocol using rational relationships for aiding cloud service ranking process. We tackle the issue of specific tasking by using a set of concepts and associated semantic rules to rank and retrieve user requirements. We promote a knowledge engineering approach for natural language processing by using terms and conditions in translating human sentences to machine readable language. Finally, we implement our system using 30 SMEs as a pivotal study. We prove that the use of semantic rules within an ontology can tackle the issue of specific tasking.