Computer Engineering

Automated Fruit Grading System is challenging task in ICT in agriculture. This paper presents a new fruit grading system using artificial neural networks (ANN). The classification systems are based on features extracted from fruits obtained from image processing feature extraction methods. Two different types of classifiers as multi-layer perceptions (MLP) with back propagation and radial basis function (RBF) networks. The objective of this study are to define a optimistic set of external quality features of fruits from the primitive features like color, texture and shape of fruits and to examine the effectiveness of the model. In this experiment, the performance of the system achieved 87.5% and 91.1% respectively for MLP with back propagation and RBF.

The volume calculation of MR Image segmented and estimate the affected intensity of Alzheimer’s disease is dealt with this paper. It is concerned with Voxel Based Morphometry to render the first part segmentation. The result gives an active region which further needs an estimation to justify the diagnosis. As in this case the image is in form of voxels. When properly processed, classified images can represent foundations for diagnostic purposes. A VBM - fuzzy approach was used to take advantage of Voxel Based Morphometry’s ability to fine segmentation based on voxel comparisons of GM, WM & CSF and membership degrees andfunctions of fuzzy logic, respectively. The method of VBM is done with SPM. The method is based on the spatial properties of the MR Brain image features and makes use of SPM multi-scaled representations of the image. A fuzzy classifier is created on basis of the previous segmented data. The method showed high quality classification for images of complex components in determining the intensity of Alzheimer’s and GUI is used as front end for the user comfort.

The ability to track the real-time location and movement of robots offers a range of useful applications in areas such as safety, security and the supply chain. Real-time Ultra Wide Band Location determination is a technology that uses Ultra-wideband (UWB) and aims at delivering high positional accuracy in harsh industrial environments that previously caused problems for traditional location systems due to electromagnetic interference. Ultra Wide Band systems can calculate the location of tags which are designed to be mounted on assets or worn by a person. They transmit UWB signals that are received by sensors which contain an array of antenna and ultra-wideband radio receivers. The data from these sensors combined with dedicated software uses algorithms to work out the angle of arrival (AOA) of the UWB signal from the tag which is then compared to the time difference of arrival (TDOA). This information is determined between pairs of sensors connected by timing cables. The combination of AOA and TDOA measurement technologies delivers a precise three dimensional location system that is powerful, reliable, and robust for harsh industrial environments. This research examines the use of ultra wide band technology in tracking items such as mobile robots.

In this world of emerging technology ultrasound imaging is widely used in medical field for disease diagnosis and treatment. Ultrasound imaging is a safe medical technique, due to its non-invasive nature, low cost, portable ultrasound system and the capability for forming real-time imaging. However the main problem with the ultrasound imaging is the image quality of ultrasound image which degrades by the presence of noise called as speckle. The speckle noise depends on the structure of tissue and various imaging parameter. This paper proposed the filtering techniques to reduce the speckle noise from ultrasound image and edge detection in ultrasound image. In this paper we describe a hybrid wiener-H filtering technique to remove speckle noise. This gives a better resulting ultrasound image for disease diagnosis and treatment by clinician.

In recent years, much research has been devoted to the exploration of operating systems; contrarily, few have studied the simulation of the memory bus. Given the current status of encrypted configurations, end-users famously desire the private unification of multi-processors and the transistor, which embodies the practical principles of algorithms. Kop, our new heuristic for the deployment of checksums, is the solution to all of these challenges.

Occurrences of certain defects are inevitable despite meticulous planning, proper process control, and good documentation during software development. Such defects lead to quality degradation which can be the underlying reason for failure. Conscious efforts are required to control and minimize software engineering defects, but this costs money, time and resources. Predicting defective software module facilitates maintenance and corrective measures. In Software Defect Prediction (SDP), predictive model estimation is based on code attributes to assess software modules containing errors likelihood. This study proposes improved activation function based on cubic spline for Multi-Layer Perceptron Neural Networks (MLPNN) to classify SDP. The proposed method is evaluated using KC1 dataset.

Electronic mail is a powerful means of communication. It is fast become a primary means of communication. It is however different from other means of communication, for instance the dimensions differ in: speed, permanence of the message; cost of distribution, deliverance to individuals and groups; an ability to filter, channel, record, and control messages. There is also a lack of ‘intimacy’ inherent in email that many believe does exist in posted letters. This can be a problem when emailing unknown persons in the hope of generating a response to a matter which relies on their taking time to ponder the query and compose a reply. We set out to examine whether posted letter would elicit more replies than emails and the rest of this paper describes the results. We conclude that posting a letter does indeed induce more replies than email.

Micro Electro Mechanical System (MEMS) controlled electromagnetic micro power generator was designed with basic cantilever structure using comsol software. This paper presents a micromechanical-system-based electromagnetic vibration-to-electrical power generator that can harvest energy from low-frequency external vibrations. The efficiency of vibration-based harvesters is proportional to excitation frequency, so the proposed generator is designed to convert low- frequency environmental vibrations to a higher frequency by employing the frequency upconversion (FupC) technique. It has been shown that the generator can effectively harvest energy from environmental vibrations of 70-150 Hz and generates 0.57-mV voltage with 0.25-nW power from a single cantilever by up converting the input vibration frequency of 95 Hz-2 kHz. The fabricated generator size is 8.5*7*2.5 mm^3,and a total of 20 serially connected cantilevers have been used to multiply the generated voltage and power. The generator demonstrated in this paper is designed for the proof of concept, and the power and voltage levels can further be increased by increasing the number of cantilevers or coil turns.

Educational blogs are used to allow educators to interact with students from any location. Teachers sharing ideas through blogs is a promising way to communicate teaching ideas and allows multicultural teaching methods to be disseminated. The blogs that tell the educator how to use technology within the classroom are informative as many teachers would have no idea about such teaching methods. Research based blogs allow users to see different educational topics discussed in more depth. This paper provides an overview of the impact blogs are having in an educational context.

Overlay text brings important semantic clues in video content analysis such as video information retrieval and summarization, since the content of the scene or the editor's intention can be well represented by using inserted text. Most of the previous approaches to extracting overlay text from videos are based on low-level features, such as edge, color, and texture information. However, existing methods experience difficulties in handling texts with various contrasts or inserted in a complex background. In this paper, we propose a novel framework to detect and extract the overlay text from the video scene. Based on our observation that there exist transient colors between inserted text and its adjacent background, a transition map is first generated. Then candidate regions are extracted by a reshaping method and the overlay text regions are determined based on the occurrence of overlay text in each candidate. In this paper represents comparative approach between edge and change in intensity models that are computationally fast and invariant to basic transformations like horizontal, Vertical and scaling. We demonstrate that change in intensity context can be used to detect overlay text regions. Intensity context describes all boundary points of the shape with respect to any single boundary point.