Understanding diversity is important in our inclusive society to hedge against ignorance and accommodate plural perspectives. Diversity nowadays can be observed in online social spaces. People from different backgrounds (e.g. gender, age, culture, expertise) are interacting every day around online digital objects (e.g. videos, images and web articles) leaving their social content in different format, commonly as textual comments and profiles. The social clouds around digital objects (i.e. user comments, user profiles and other metadata of digital objects) offer rich source of information about the users and their perspectives on different domains. Although, researchers from disparate disciplines have been working on understanding and measuring diversity from different perspectives, little has been done to automatically measure diversity in social clouds. This is the main objective of this research. This research proposes a semantic driven computational model to systematically represent and automatically measure diversity in a social cloud. Definitions from a prominent diversity framework and Semantic Web techniques underpin the proposed model. Diversity is measured based on four diversity indices - variety, balance, coverage and (within and across) disparity with regards to two perspectives – (a) domain, which is captured in user comments and represented by domain ontologies, and (b) user, which is captured in profiles of users who made the comments and represented by a proposed User Diversity Ontology. The proposed model is operationalised resulting in a Semantic Driven Diversity Analytics Tool (SeDDAT), which is responsible for diversity profiling based on the diversity indices. The proposed approach of applying the model is illustrated on social clouds from two social spaces - open (YouTube) and closed (Active Video Watching (AVW-Space)). The open social cloud shows the applicability of the model to generate diversity profiles of a large pool of videos (600) with thousands of users and comments. Closed social clouds of two user groups around same set of videos illustrate transferability and further utility of the model. A list of possible diversity patterns within social clouds is provided, which in turn deepen the understanding of diversity and open doors for further utilities of the diversity profiles. The proposed model is applicable in similar scenarios, such as in the social clouds around MOOCs and news articles.

Colour image watermarking has become one of the most important algorithms for copyright protection. The following paper will present an innovative scheme for watermarking blind colour images using the discrete wavelet transform (DWT), fast Walsh Hadamard transform (FWHT) and the Hessenberg decomposition as its basis. First, two-level DWT followed by FWHT are used to decompose the host image’s red channel. Next, the FWHT coefficients are split into 4× 4 non-overlapping blocks. Then, each selected block is decomposed using Hessenberg decomposition, where the first row, first column element of the upper Hessenberg matrix H is quantified to embed the watermark information. Peak signalto-noise ratio, normalized cross-correlation and structural similarity index measure are used to evaluate the feasibility and the robustness. The experimental results have demonstrated that the proposed watermarking scheme is highly invisible with PSNR> 40 dB, for several watermarked colour images, with a capacity of 4096 bits and execution time of 0.7415 s. The proposed watermarking scheme is also highly resistant to both common image processing and geometrical attacks such as filtering, JPEG2000, noise adding, cropping, scaling, blurring and sharpening, and others

Modelling diversity is especially valuable in soft skills learning, where contextual awareness and understanding of different perspectives are crucial. This paper presents an application of a diversity analytics pipeline to generate domain diversity profiles for learners as captured in their comments while watching videos for learning a soft skill. The datasets for analysis were collected from a series of studies on learning presentation skills with Active Video Watching system (AVW-Space). Two user studies (with 37 postgraduates and 140 undergraduates, respectively) were compared. The learners’ diversity and personal profiles are used to further understand and highlight any notable patterns about their domain coverage on presentation skills.

Diversity has been the subject of study in various disciplines from biology to social science and computing. Respecting and utilising the diversity of the population is increasingly important to broadening knowledge. This paper describes a pipeline for diversity profiling of a pool of text in order to understand its coverage of an underpinning domain. The application is illustrated by using a domain ontology on presentation skills in a case study with 38 postgraduates who made comments while learning pitch presentations with the Active Video Watching system (AVW-Space). The outcome shows different patterns of coverage on the domain by the comments in each of the eight videos.

Abstract—Early detection of malignant melanoma skin cancer is crucial for treating the disease and saving lives. Many computerized techniques have been reported in the literature to diagnose and classify the disease with satisfactory skin cancer detection performance. However, reducing the false detection rate is still challenging and preoccupying because false positives trigger the alarm and require intervention by an expert pathologist for further examination and screening. In this paper, an automatic skin cancer diagno- sis system that combines different textural and color features is proposed. New textural and color features are used in a bag-of-features approach for efficient and accurate detection. We particularly claim that the Histogram of Gradients (HG) and the Histogram of Lines (HL) are more suitable for the analysis and clas- sification of dermoscopic and standard skin images than the conventional Histogram of Oriented Gradient (HOG) and the Histogram of Oriented Lines (HOL), respectively. The HG and HL are bagged separately us- ing a codebook for each and then combined with other bagged color vector angles and Zernike moments to exploit the color information. The overall system has been assessed through intensive experiments using different classifiers on a dermoscopic image dataset and another standard dataset. Experimental results have shown the superiority of the proposed system over state-of-the-art techniques.

The current paper proposes a novel scheme for non-blind watermarking of images, making use of discrete wavelet transform (DWT), discrete time Fourier transform (DTFT), as well as singular value decomposition, or SVD. During the process of embedding, 1-level DWT is used to decompose the host image into its various frequency sub-bands. After this, the high-frequency sub band receives an application of DTFT. This is followed then by SVD, after which the watermark becomes embedded into the now-transformed host image's singular matrix. Then, the inverses of 1-level DWT, DTFT and SVD are applied in order to obtain a watermarked final image. This paper evaluates the performance of the proposed method of watermarking against a number of attacks, including sharpening, salt and pepper noise, AWGN, gamma correction, histogram equalisation, flipping and cropping. Results obtained during experiments have found that the scheme as proposed does provide high levels of robustness and imperceptibility against various signal processing attacks.

This paper presents a computational model for measuring diversity in terms of variety, balance and disparity. This model is informed by the Stirling’s framework for understanding diversity from social science and underpinned by semantic techniques from computer science. A case study in learning is used to illustrate the application of the model. It is driven by the desire to broaden learners’ perspectives in an increasingly diverse and inclusive society. For example, interpreting body language in a job interview may be influenced by the different background of observers. With the explosion of digital objects on social platforms, selecting the appropriate ones for learning can be challenging and time consuming. The case study uses over 2000 annotated comments from 51 YouTube videos on job interviews. Diversity indicators are produced based on the comments for each video, which in turn facilitate the ranking of the videos according to the degree of diversity in the comments for the selected domain.

Abstract—With a rapidly increasing incidence of melanoma

skin cancer, there is a need for decision support systems to

detect it in its early stages, which would lead to better decisions

in treating it successfully. However, developing such systems is

still a challenging task for researchers. Several Computer Aided-

Diagnosis (CAD) systems have been proposed in the last two

decades to increase the accuracy of melanoma detection. Image

feature extraction is a critical step in differentiating between

melanoma and normal skin lesions. In this paper, we propose

to improve a bag-of-words approach by combining features

consisting of the color histogram and first order moments with the

Histogram of Oriented Gradients (HOG). Experimental results

show that the proposed technique significantly improves the

detection accuracy, with an average sensitivity of 91% and

specificity of 85%. The proposed system was validated on a

dataset of 200 medically annotated images (40 melanomas and

160 non-melanomas) obtained from the database of the Hospital

Pedro Hispano. 

Abstract— Diagnosing skin cancer in its early stages is a challenging task for dermatologists given the fact that the chance for a patient’s survival is higher and hence the process of analyzing skin images and making decisions should be time efficient. Therefore, diagnosing the disease using automated and computerized systems has nowadays become essential. This paper proposes an efficient system for skin cancer detection on dermoscopic images. It has been shown that the statistical characteristics of the pigment network, extracted from the dermoscopic image, could be used as efficient discriminating features for cancer detection. The proposed system has been assessed on a dataset of 200 dermoscopic images of the ‘Hospital Pedro Hispano’ [1] and the results of cross-validation have shown high detection accuracy.

The aim was to build an educational website to teach preschoolers aged 2-5 years old some skills; related to reading, writing and pronouncing alphabet, numbers, animal names, shapes and other common words. Also, assist parents and teachers by providing digital game-based learning environment to “edutainment” the children. The study illustrated the usability and usefulness of this educational game-based website and provided future insights for researchers and developers to take into account in future work.