Abstract
Current state-of-the-art medical image segmentation methods require high quality datasets to obtain good performance. However, medical specialists often disagree on diagnosis, hence, datasets contain contradictory annotations. This, in turn, leads to difficulties in the optimization process of Deep Learning models and hinder performance. We propose a method to estimate uncertainty in Convolutional Neural Network (CNN) segmentation models, that makes the training of CNNs more robust to contradictory annotations. In this work, we model two types of uncertainty, heteroscedastic and epistemic, without adding any additional supervisory signal other than the ground-truth segmentation mask. As expected, the uncertainty is higher closer to vessel boundaries, and on top of thinner and less visible vessels where it is more likely for medical specialists to disagree. Therefore, our method is more suitable to learn from datasets created with heterogeneous annotators. We show that there is a correlation between the uncertainty estimated by our method and the disagreement in the segmentation provided by two different medical specialists. Furthermore, by explicitly modeling the uncertainty, the Intersection over Union of the segmentation network improves 5.7 percentage points.

Abstract
Vulnerability scanning tools can help secure the computer networks of organisations. Triggered by the release of the Tsunami vulnerability scanner by Google, the authors analysed and compared the commonly used, free-to-use vulnerability scanners. The performance, accuracy and precision of these scanners are quite disparate and vary accordingly to the target systems. The computational, memory and network resources required be these scanners also differ. We present a recent and detailed comparison of such tools that are available for use by organisations with lower resources such as small and medium-sized enterprises.

Abstract
This paper discusses a novel collective sensing approach using autonomous sensors specially designed to monitor gas leaks and search for gas sources. The proposed collective behavior aims to improve the gas-source search by sharing information between mobile sensors and reducing the risks associated with gas leakage. The group acts as a composite sensor that can move independently to search for an optimal sensing zone. The autonomous searching behavior is bio-inspired by colonies of bacteria that continuously seek energy sources throughout their existence. Each sensor makes its own autonomous search decision, considering the group sense, to move in the direction of a better energy source. The collective approach is based on autonomous agents sharing information to achieve a collective sense of gas perception and utilizes more intelligent searching. The method is evaluated in a cyber-physical system specially developed to safely experiment with gases and mobile sensors while reproducing the realistic dynamic behavior of the gas. Experiments are performed to clarify the collective gas-sensing contributions, and the gas search is compared through multiple mobile sensors with and without collective sensing. The proposed approach is evaluated in an unhealthy environment to elucidate its effectiveness. In addition to presenting the related differences between collective and individual sensory approaches, this work contributes with analyzes of the scalability of mobile gas sensing systems. This work also contributed as a simulated semi-physical experimental system to test algorithms' performance before applying it to practice.

Abstract

Abstract
Computer network attacks are vast and negatively impact the infrastructure and its applicational services. From a cyber offensive and defensive perspective, there are a plethora of tools to craft and inject customized malicious packets in the network and exploit operating systems and application vulnerabilities. Those tools are however hard to operate by practitioners with less knowledge on networking fundamentals and students in the early stage of their studies. This paper proposes an easy-to-use application tool that can produce customized Denial of Service (DoS) and spoofing attacks. It was developed in Python and takes advantage of scapy library to process and inject network packets. A set of experiments was made, and the results obtained show the efficiency and accuracy of the attacks, by impairing the proper functioning of the victim's machines.

Abstract
Purpose This study aims to understand why young people are interested in buying frugal innovations. Design/methodology/approach Data were collected with a survey administered to 534 university students enrolled in various fields of study (e.g. sciences, technology, economics and fine arts). Using the Tata Nano car as an example of frugal innovation, a model based on the unified theory of acceptance and use of technology was developed using partial least squares structural equation modeling. Findings The model's results reveal that effort expectancy, performance expectancy and facilitating conditions are critical factors that explain university students' intention to buy Tata Nano. Originality/value Although frugal innovations are often introduced first in developing countries, frugal innovations could be highly relevant to users in developed nations as these innovations can provide market opportunities in terms of cost-conscious, relatively low-income and sustainability-conscious consumers.