Abstract
Due to inevitable differences between the data used for training modern CAD systems and the data encountered when they are deployed in clinical scenarios, the ability to automatically assess the quality of predictions when no expert annotation is available can be critical. In this paper, we propose a new method for quality assessment of retinal vessel tree segmentations in the absence of a reference ground-truth. For this, we artificially degrade expert-annotated vessel map segmentations and then train a CNN to predict the similarity between the degraded images and their corresponding ground-truths. This similarity can be interpreted as a proxy to the quality of a segmentation. The proposed model can produce a visually meaningful quality score, effectively predicting the quality of a vessel tree segmentation in the absence of a manually segmented reference. We further demonstrate the usefulness of our approach by applying it to automatically find a threshold for soft probabilistic segmentations on a per-image basis. For an independent state-of-the-art unsupervised vessel segmentation technique, the thresholds selected by our approach lead to statistically significant improvements in F1-score (+2.67%) and Matthews Correlation Coefficient (+3.11%) over the thresholds derived from ROC analysis on the training set. The score is also shown to correlate strongly with F1 and MCC when a reference is available.

Abstract
In this work we study a way to explore and extract more information from data sets with a hierarchical tree structure. We propose that any statistical study on this type of data should be made by group, after clustering. In this sense, the most adequate approach is to use the Mahalanobis-Wasserstein distance as a measure of similarity between the cases, to carry out clustering or unsupervised classification. This methodology allows for the clustering of cases, as well as the identification of their profiles, based on the distribution of all the variables that characterises each subject associated with each case. An application to a set of teenagers' interviews regarding their habits of communication is described. The interviewees answered several questions about the kind of contacts they had on their phone, Facebook, email or messenger as well as the frequency of communication between them. The results indicate that the methodology is adequate to cluster this kind of data sets, since it allows us to identify and characterise different profiles from the data. We compare the results obtained with this methodology with the ones obtained using the entire database, and we conclude that they may lead to different findings.

Abstract
Teaching/educating patients with end stage renal disease (ESRD) and identifying their self-care behaviors for vascular network preservation are very important. However, the self-care behaviors regularly performed by patients are still unknown. We compared self-care behaviors for vascular network preservation performed by patients who are/are not followed-up by the nephrologist. The study design was a prospective, observational and comparative study. Inclusion criteria were as follows: ESRD patients (at stages 4 or 5); at least 18 years old; in pre-dialysis with at least a 6-month follow-up period by the nephrologist or who started dialysis in emergency and were not followed-up by the nephrologist; with no memory problems; and medically stable. Primary outcome was the frequency of self-care behaviors for vascular network preservation. Secondary outcome was the comparison between self-care behaviors by ESRD patients who were/were not followed-up by the nephrologist. The study involved 145 patients, 64.1% were female, the mean age was 69.5 years and the self-care behaviors mean score was 36.8% (with a SD of 39.8%). The number of patients followed-up and not followed-up by the nephrologist was 109 (group 1) and 36 (group 2), respectively. Social characteristics were similar in the two groups (P > 0.05). The mean self-care behaviors were 29.4% and 59.2% in groups 1 and 2, respectively (P = 0.000). Patients performed self-care behaviors for vascular network preservation with a relatively low frequency (the mean score was 36.8% only). Patients not followed by the nephrologist performed self-care behaviors more often than those who were followed (59.2% vs. 29.4% respectively, P = 0.000).

Abstract
The increasing need for more adaptive production environments is a big motivator for the adoption of agent-based technologies in industrial systems, as they provide better mechanisms for handling dynamically and intelligently various kinds of production disturbances. Unlike with the utilization of most conventional automation languages, the use of agents enables, in an easy way, the setup of dynamic and autonomous adaptive processes to handle large and complex engineering system functions and interactions. Agent-technologies in cyber-physical systems contexts require at some point integration with automation controllers. However, most commonly available and used agent system implementations in the industry were not designed for hard real-time control use cases, and do not utilize real-time operating systems or dedicated hardware. Hence, they cannot match the hard-real-time performance of automation controllers. This work provides some insights on the performance that can be achieved with agent-based approaches that integrate with low-level automation system functions. It considers the performance of the agent-based practices in light of non-real-time dedicated hardware or operating systems. The results show that agents are well suited for the majority of soft-real-time control applications.

Abstract
Organizations of the agro-industrial sector, are now increasingly investing in the development of technological systems that allow the computerization of all its processes. Recently the methods and techniques of computer vision have been widely used for monitoring and inspection during the production and harvesting, allowing detect problems early and thus, improve the quality of products. In the field of mushroom production one of the most important aspects, and perhaps most prevalent, is to be able to predict its production. To this end it is proposed an Intelligent System Mushroom Harvest Forecast (SIPCC), based on techniques and methods of computer vision and Artificial Neural Networks (ANN). This paper presents an architecture of a SIPCC functional and technical level, complemented with the analysis and presentation of data demonstrating its viability.

Abstract