Machine vision and planning applied on assembby line of multiple products

This article describes the use of Artificial Intelligence (IA) techniques applied in cells of a manufacturing system. Machine Vision was used to identify pieces and their positions of two different products to be assembled in the same productive line. This information is given as input for an IA planner embedded in the manufacturing system. Therefore, initial and final states are sent automatically to the planner capable to generate assembly plans for a robotic cell, in real time.

Color vision impairment in type 2 diabetes assessed by the D-15d test and the Cambridge Colour Test

Color vision impairment emerges at early stages of diabetes mellitus type 2 (DM2) and may precede diabetic retinopathy or the appearance of vascular alterations in the retina. The aim of the present study was to compare the evaluation of the color vision with two different tests - the Lanthony desaturated D-15d test (a traditional color arrangement test), and the Cambridge Colour Test (CCT) (a computerized color discrimination test) - in patients diagnosed with DM2 without clinical signs of diabetic retinopathy (DR), and in sex- and age-matched control groups. Both color tests revealed statistically significant differences between the controls and the worst eyes of the DM2 patients. In addition, the degree of color vision impairment diagnosed by both tests correlated with the disease duration. The D-15d outcomes indicated solely tritan losses. In comparison, CCT outcomes revealed diffuse losses in color discrimination: 13.3% for best eyes and 29% for worst eyes. In addition, elevation of tritan thresholds in the DM2 patients, as detected by the Trivector subtest of the CCT, was found to correlate with the level of glycated hemoglobin. Outcomes of both tests confirm that subclinical losses of color vision are present in DM2 patients at an early stage of the disease, prior to signs of retinopathy. Considering the advantages of the CCT test compared to the D-15d test, further studies should attempt to verify and/or improve the efficiency of the CCT test.

Using the Hard, Randy, and Rittler Test to Evaluate Color Vision in Capuchins (Cebus libidinosus)

The identification of color vision types in primates is fundamental to understanding the evolution and biological function of color perception. The Hard, Randy, and Rittler (HRR) pseudoisochromatic test categorizes human color vision types successfully. Here we provide an experimental setup to employ HRR in a nonhuman primate, the capuchin (Cebus libidinosus), a platyrrhine with polymorphic color vision. The HRR test consists of plates with a matrix composed of gray circles that vary in size and brightness. Differently colored circles form a geometric shape (X, O, or Delta) that is discriminated visually from the gray background pattern. The ability to identify these shapes determines the type of dyschromatopsy (deficiency in color vision). We tested six capuchins in their own cages under natural sunlight. The subjects chose between two HRR plates in each trial: one with the gray pattern only and the other with a colored shape, presented on the left or right side at random. We presented the test 40 times and calculated the 95 % confidence limits for chance performance based on the binomial test. We also genotyped all subjects for exons 3 and 5 of the X-linked opsin genes. The HRR test diagnosed two subjects as protan dichromats (missing or defective L-cone), three as deutan dichromats (missing or defective M-cone), and one female as trichromat. Genetic analysis supported the behavioral data for all subjects. These findings show that the HRR test can be applied to diagnose color vision in nonhuman primates.

Long-Term Occupational Exposure to Organic Solvents Affects Color Vision, Contrast Sensitivity and Visual Fields

The purpose of this study was to evaluate the visual outcome of chronic occupational exposure to a mixture of organic solvents by measuring color discrimination, achromatic contrast sensitivity and visual fields in a group of gas station workers. We tested 25 workers (20 males) and 25 controls with no history of chronic exposure to solvents (10 males). All participants had normal ophthalmologic exams. Subjects had worked in gas stations on an average of 9.6 +/- 6.2 years. Color vision was evaluated with the Lanthony D15d and Cambridge Colour Test (CCT). Visual field assessment consisted of white-on-white 24-2 automatic perimetry (Humphrey II-750i). Contrast sensitivity was measured for sinusoidal gratings of 0.2, 0.5, 1.0, 2.0, 5.0, 10.0 and 20.0 cycles per degree (cpd). Results from both groups were compared using the Mann-Whitney U test. The number of errors in the D15d was higher for workers relative to controls (p<0.01). Their CCT color discrimination thresholds were elevated compared to the control group along the protan, deutan and tritan confusion axes (p<0.01), and their ellipse area and ellipticity were higher (p<0.01). Genetic analysis of subjects with very elevated color discrimination thresholds excluded congenital causes for the visual losses. Automated perimetry thresholds showed elevation in the 9 degrees, 15 degrees and 21 degrees of eccentricity (p<0.01) and in MD and PSD indexes (p<0.01). Contrast sensitivity losses were found for all spatial frequencies measured (p<0.01) except for 0.5 cpd. Significant correlation was found between previous working years and deutan axis thresholds (rho = 0.59; p<0.05), indexes of the Lanthony D15d (rho = 0.52; p<0.05), perimetry results in the fovea (rho = -0.51; p<0.05) and at 3, 9 and 15 degrees of eccentricity (rho = -0.46; p<0.05). Extensive and diffuse visual changes were found, suggesting that specific occupational limits should be created.

Contribution of Vision and Proprioception to the Precision of Reaching Movements

Ren and colleagues (2006) found that saccades to visual targets became less accurate when somatosensory information about hand location was added, suggesting that saccades rely mainly on vision. We conducted two kinematic experiments to examine whether or not reaching movements would also show such strong reliance on vision. In Experiment 1, subjects used their dominant right hand to perform reaches, with or without a delay, to an external visual target or to their own left fingertip positioned either by the experimenter or by the participant. Unlike saccades, reaches became more accurate and precise when proprioceptive information was available. In Experiment 2, subjects reached toward external or bodily targets with differing amounts of visual information. Proprioception improved performance only when vision was limited. Our results indicate that reaching movements, unlike saccades, are improved rather than impaired by the addition of somatosensory information.

Development of an X-ray spectrometric system and feasibility tests of Silicon Drift Detector for medical and space applications

The thesis work concerns X-ray spectrometry for both medical and space applications and is divided into two sections. The first section addresses an X-ray spectrometric system designed to study radiological beams and is devoted to the optimization of diagnostic procedures in medicine. A parametric semi-empirical model capable of efficiently reconstructing diagnostic X-ray spectra in 'middle power' computers was developed and tested. In addition, different silicon diode detectors were tested as real-time detectors in order to provide a real-time evaluation of the spectrum during diagnostic procedures. This project contributes to the field by presenting an improved simulation of a realistic X-ray beam emerging from a common X-ray tube with a complete and detailed spectrum that lends itself to further studies of added filtration, thus providing an optimized beam for different diagnostic applications in medicine. The second section describes the preliminary tests that have been carried out on the first version of an Application Specific Integrated Circuit (ASIC), integrated with large area position-sensitive Silicon Drift Detector (SDD) to be used on board future space missions. This technology has been developed for the ESA project: LOFT (Large Observatory for X-ray Timing), a new medium-class space mission that the European Space Agency has been assessing since February of 2011. The LOFT project was proposed as part of the Cosmic Vision Program (2015-2025).

The posterior parietal cortex: a bridge between vision and action

The present work takes into account three posterior parietal areas, V6, V6A, and PEc, all operating on different subsets of signals (visual, somatic, motor). The work focuses on the study of their functional properties, to better understand their respective contribution in the neuronal circuits that make possible the interactions between subject and external environment. In the caudalmost pole of parietal lobe there is area V6. Functional data suggest that this area is related to the encoding of both objects motion and ego-motion. However, the sensitivity of V6 neurons to optic flow stimulations has been tested only in human fMRI experiments. Here we addressed this issue by applying on monkey the same experimental protocol used in human studies. The visual stimulation obtained with the Flow Fields stimulus was the most effective and powerful to activate area V6 in monkey, further strengthening this homology between the two primates. The neighboring areas, V6A and PEc, show different cytoarchitecture and connectivity profiles, but are both involved in the control of reaches. We studied the sensory responses present in these areas, and directly compared these.. We also studied the motor related discharges of PEc neurons during reaching movements in 3D space comparing also the direction and depth tuning of PEc cells with those of V6A. The results show that area PEc and V6A share several functional properties. Area PEc, unlike V6A, contains a richer and more complex somatosensory input, and a poorer, although complex visual one. Differences emerged also comparing the motor-related properties for reaches in depth: the incidence of depth modulations in PEc and the temporal pattern of modulation for depth and direction allow to delineate a trend among the two parietal visuomotor areas.

Requirements definition and testing of the new version of multi-mission generic monitoring system (GEMS) v 4.4 for operations at Eumetsat

EUMETSAT (www.eumetsat.int) e’ l’agenzia europea per operazioni su satelliti per monitorare clima, meteo e ambiente terrestre. Dal centro operativo situato a Darmstadt (Germania), si controllano satelliti meteorologici su orbite geostazionarie e polari che raccolgono dati per l’osservazione dell’atmosfera, degli oceani e della superficie terrestre per un servizio continuo di 24/7. Un sistema di monitoraggio centralizzato per programmi diversi all’interno dell’ambiente operazionale di EUMETSAT, e’ dato da GEMS (Generic Event Monitoring System). Il software garantisce il controllo di diverse piattaforme, cross-monitoring di diverse sezioni operative, ed ha le caratteristiche per potere essere esteso a future missioni. L’attuale versione della GEMS MMI (Multi Media Interface), v. 3.6, utilizza standard Java Server Pages (JSP) e fa uso pesante di codici Java; utilizza inoltre files ASCII per filtri e display dei dati. Conseguenza diretta e’ ad esempio, il fatto che le informazioni non sono automaticamente aggiornate, ma hanno bisogno di ricaricare la pagina. Ulteriori inputs per una nuova versione della GEMS MMI vengono da diversi comportamenti anomali riportati durante l’uso quotidiano del software. La tesi si concentra sulla definizione di nuovi requisiti per una nuova versione della GEMS MMI (v. 4.4) da parte della divisione ingegneristica e di manutenzione di operazioni di EUMETSAT. Per le attivita’ di supporto, i test sono stati condotti presso Solenix. Il nuovo software permettera’ una migliore applicazione web, con tempi di risposta piu’ rapidi, aggiornamento delle informazioni automatico, utilizzo totale del database di GEMS e le capacita’ di filtri, insieme ad applicazioni per telefoni cellulari per il supporto delle attivita’ di reperibilita’. La nuova versione di GEMS avra’ una nuova Graphical User Interface (GUI) che utilizza tecnologie moderne. Per un ambiente di operazioni come e’ quello di EUMETSAT, dove l’affidabilita’ delle tecnologie e la longevita’ dell’approccio scelto sono di vitale importanza, non tutti gli attuali strumenti a disposizione sono adatti e hanno bisogno di essere migliorati. Allo stesso tempo, un’ interfaccia moderna, in termini di visual design, interattivita’ e funzionalita’, e’ importante per la nuova GEMS MMI.

Deep learning for computer vision: a comparison between convolutional neural networks and hierarchical temporal memories on object recognition tasks

In recent years, Deep Learning techniques have shown to perform well on a large variety of problems both in Computer Vision and Natural Language Processing, reaching and often surpassing the state of the art on many tasks. The rise of deep learning is also revolutionizing the entire field of Machine Learning and Pattern Recognition pushing forward the concepts of automatic feature extraction and unsupervised learning in general. However, despite the strong success both in science and business, deep learning has its own limitations. It is often questioned if such techniques are only some kind of brute-force statistical approaches and if they can only work in the context of High Performance Computing with tons of data. Another important question is whether they are really biologically inspired, as claimed in certain cases, and if they can scale well in terms of "intelligence". The dissertation is focused on trying to answer these key questions in the context of Computer Vision and, in particular, Object Recognition, a task that has been heavily revolutionized by recent advances in the field. Practically speaking, these answers are based on an exhaustive comparison between two, very different, deep learning techniques on the aforementioned task: Convolutional Neural Network (CNN) and Hierarchical Temporal memory (HTM). They stand for two different approaches and points of view within the big hat of deep learning and are the best choices to understand and point out strengths and weaknesses of each of them. CNN is considered one of the most classic and powerful supervised methods used today in machine learning and pattern recognition, especially in object recognition. CNNs are well received and accepted by the scientific community and are already deployed in large corporation like Google and Facebook for solving face recognition and image auto-tagging problems. HTM, on the other hand, is known as a new emerging paradigm and a new meanly-unsupervised method, that is more biologically inspired. It tries to gain more insights from the computational neuroscience community in order to incorporate concepts like time, context and attention during the learning process which are typical of the human brain. In the end, the thesis is supposed to prove that in certain cases, with a lower quantity of data, HTM can outperform CNN.

Outcomes following three-line vision loss during treatment of neovascular age-related macular degeneration: subgroup analyses from MARINA and ANCHOR

This study aims to assess the impact of continued ranibizumab treatment for neovascular age-related macular degeneration on patients from the MARINA and ANCHOR randomised clinical studies who lost ≥ 3 lines of best-corrected visual acuity (BCVA) at any time during the first year of treatment.