Search

In this article we show the change in paradigm occurred in color constancy algorithms: from a pre-processing step in image understanding, to the exploitation of image understanding and computer vision results and techniques. Since color constancy is an ill-posed problem, we give an overview of the assumptions on which classical color constancy algorithms are based in order to solve it. Then, we chronologically review the color constancy algorithms that exploit results and techniques borrowed from the image understanding research field in order to exploit assumptions that could be met in a larger number of images.

Illumination estimation is the essential step of computational color constancy, one of the core parts of various image processing pipelines of modern digital cameras. Having an accurate and reliable illumination estimation is important for reducing the illumination influence on the image colors. To motivate the generation of new ideas and the development of new algorithms in this field, two challenges on illumination estimation were conducted. The main advantage of testing a method on a challenge over testing it on some of the known datasets is the fact that the ground- truth illuminations for the challenge test images are unknown up until the results have been submitted, which prevents any potential hyperparameter tuning that may be biased. The First illumination estimation challenge (IEC#1) had only a single task, global illumination estimation. The second illumination estimation challenge (IEC#2) was enriched with two additional tracks that encompassed indoor and two-illuminant illumination estimation. Other main features of it are a new large dataset of images (about 5000) taken with the same camera sensor model, a manual markup accompanying each image, diverse content with scenes taken in numerous countries under a huge variety of illuminations extracted by using the SpyderCube calibration object, and a contest- like markup for the images from the Cube++ dataset. This article focuses on the description of the past two challenges, algorithms which won in each track, and the conclusions that were drawn based on the results obtained during the first and second challenge that can be useful for similar future developments.

Bakurov, I., Buzzelli, M., Castelli, M., Schettini, R., & Vanneschi, L. (2020). Parameters optimization of the Structural Similarity Index. In London Imaging Meeting 2020: Future Colour Imaging (1 ed., Vol. 2020, pp. 19-23). (London Imaging Meeting). https://doi.org/10.2352/issn.2694-118X.2020.LIM-13 We exploit evolutionary computation to optimize the handcrafted Structural Similarity method (SSIM) through a datadriven approach. We estimate the best combination of luminance, contrast and structure components, as well as the sliding window size used for processing, with the objective of optimizing the similarity correlation with human-expressed mean opinion score on a standard dataset. We experimentally observe that better results can be obtained by penalizing the overall similarity only for very low levels of luminance similarity. Finally, we report a comparison of SSIM with the optimized parameters against other metrics for full reference quality assessment, showing superior performance on a different dataset. publishersversion published

Background and objective: Autism spectrum disorder (ASD) is a complex neurodevelopmental disorder for which no treatments exist. Fragile X syndrome (FXS) is the most common form of inherited mental retardation and the most frequent monogenic cause of ASD. Given the lack of pharmacological treatments for ASD, increasing interest is devoted to non-pharmacological approaches, including dietary interventions. Omega-3 polyunsaturated fatty acids (PUFAs) are critical for neurobehavioraldevelopment. This study had two aims: 1. To validatethe recently developed Fmr1-Δexon 8 rat model of FXS; 2. To assess the impact of omega-3 PUFAs dietary supplementation during pregnancy and lactation on the altered behavior displayed by Fmr1-Δexon 8 rats. Methods: Female Fmr1-Δexon 8 and wild-type Sprague–Dawley rats were fed with either an omega-3 PUFAs enriched diet or with an isocaloric control diet during pregnancy and lactation. Behavioral experiments were carried out on the infant (Postnatal days (PNDs) 9 and 13), juvenile (PND 35) and adult (PND 90) male offspring. Results:Fmr1-Δexon 8 pups showed hypolocomotion, reduced ultrasonic vocalizations (USVs) emission and impaired social discrimination compared to wild-type controls. Juvenile and adult Fmr1-Δexon 8 rats showed deficits in the social and cognitive domains, that were counteracted by perinatal omega-3 PUFAs supplementation. Conclusion: Our results support the validity of the Fmr1-Δexon 8 rat model to mimic key autistic-like features and support an important role of omega-3 PUFAs during of neurodevelopment. Although the mechanisms underlying the beneficial effects of omega-3 PUFAs supplementation in ASD needs to be clarified, this dietary intervention holds promise to mitigate core and comorbid autistic features.

Methylenedioxypyrovalerone (MDPV) is the most popular synthetic cathinone found in products marketed as 'bath salts', widely abused among teenagers and young adults. Synthetic cathinones have pharmacological effects resembling those of psychostimulants, which are known to disrupt a variety of social behaviors. However, despite the popular use of MDPV by young people in social contexts, information about its effects on social behavior is scarce. To investigate the impact of MDPV on social behavior at young age, and the underlying neurobehavioral mechanisms, we focused on social play behavior. Social play behavior is the most characteristic social behavior displayed by young mammals and it is crucial for neurobehavioral development. Treatment with MDPV reduced social play behavior in both juvenile and young adult male rats, and its play-suppressant effect was subject to tolerance but not sensitization. As the behavioral effects of MDPV have been ascribed to dopaminergic and noradrenergic neurotransmission, and given the role of these neurotransmitters in social play, we investigated the involvement of dopamine and noradrenaline in the play-suppressant effects of MDPV. The effects of MDPV on social play were blocked by either the α2 adrenoceptor antagonist RX821002 or the dopamine receptor antagonist flupenthixol, given alone or together at sub-effective doses. In sum, MDPV selectively suppresses the most vigorous social behavior of developing rats through both noradrenergic and dopaminergic mechanisms. This study provides important preclinical evidence of the deleterious effects of MDPV on social behavior, and as such increases our understanding of the neurobehavioral effects of this popular cathinone.

Bakurov, I., Buzzelli, M., Schettini, R., Castelli, M., & Vanneschi, L. (2022). Structural similarity index (SSIM) revisited: A data-driven approach. Expert Systems with Applications, 189, 1-19. [116087]. [Advanced online publication on 27 October 2021]. https://doi.org/10.1016/j.eswa.2021.116087--------------Funding Information: This work was supported by national funds through the FCT (Funda??o para a Ci?ncia e a Tecnologia), Portugal by the projects GADgET (DSAIPA/DS/0022/2018), BINDER (PTDC/CCIINF/29168/2017), and AICE (DSAIPA/DS/0113/2019). Mauro Castelli acknowledges the financial support from the Slovenian Research Agency, Slovenia (research core funding no. P5-0410). Several contemporaneous image processing and computer vision systems rely upon the full-reference image quality assessment (IQA) measures. The single-scale structural similarity index (SS-SSIM) is one of the most popular measures, and it owes its success to the mathematical simplicity, low computational complexity, and implicit incorporation of Human Visual System’s (HVS) characteristics. In this paper, we revise the original parameters of SSIM and its multi-scale counterpart (MS-SSIM) to increase their correlation with subjective evaluation. More specifically, we exploit the evolutionary computation and the swarm intelligence methods on five popular IQA databases, two of which are dedicated distance-changed databases, to determine the best combination of parameters efficiently. Simultaneously, we explore the effect of different scale selection approaches in the context of SS-SSIM. The experimental results show that with a proper fine-tuning (1) the performance of SS-SSIM and MS-SSIM can be improved, in average terms, by 8% and by 3%, respectively, (2) the SS-SSIM after the so-called standard scale selection achieves similar performance as if applying computationally more expensive state-of-the-art scale selection methods or MS-SSIM; moreover, (3) there is evidence that the parameters learned on a given database can be successfully transferred to other (previously unseen) databases; finally, (4) we propose a new set of reference parameters for SSIM’s variants and provide their interpretation. authorsversion published

This article aims to illustrate an innovative and absolutely functional method of assessing the fitness of a tennis player, based on the registration of an incremental metabolic effort, under attentive aegis.

Highlights • Health professional students can effectively implement a contact tracing initiative • A replicable workflow and onboarding system aids in efficient contact tracing • Testing capacity, community buy-in, and sparse resources can impact contact tracing, Background The Centers for Disease Control and Prevention recommends aggressive contact tracing to control the COVID-19 pandemic. In this work, we (1) describe the development of a COVID-19 contact tracing initiative that includes medical, nursing, and public health students, and is led by clinicians and infectious disease epidemiologists within our health system, and, (2) articulate process steps for contact tracing including workflows and telephone scripts, and, (3) highlight the key challenges and strategies to overcome these challenges. Methods A single academic institution-based contact tracing initiative was rapidly scaled to 110 health professional students, four physicians, two epidemiologists, and a research team. Following training, students called patients who were COVID-19 positive and the individuals they were in contact with to ensure proper isolation and quarantine measures. Students also assisted those who faced barriers to quarantine. Implications In total, between March 24 and May 28 – this initiative completed contact tracing for 536 confirmed cases, which resulted in the identification of 953 contacts. We aim to disseminate this process, including telephone scripts and workflow, to other health systems for use in their initiatives to respond to the COVID-19 pandemic and future public health emergencies.

We present a method for illuminant estimation that exploits a generative adversarial network architecture to generate a spatially-varying illuminant map. This map is then transformed by consensus into a global illuminant estimation, in the form of a single RGB triplet. To this end, different consensus strategies are designed and compared in this paper. The best solution won second place in the 2nd International Illumination Estimation Challenge, specifically for the indoor track.

Recently, Computer Vision based image analysis techniques have attracted a lot of attention because they are used to develop automatic dietary monitoring applications. Food recognition is a quite challenging task: it is a non-rigid object, and is characterized by intrinsic high iter- and intra-class variability. The proper design of a food recognition system based on Computer Vision should contain several analysis stages. This paper reports on the most recent solutions in the field of automatic food recognition using computer vision developed at the Imaging and Vision Laboratory in the last 12 years. We present and discuss the main solutions developed and results achieved for food localization, segmentation, recognition and analysis. Food localization and segmentation aim at identifying the regions in the image corresponding to food items, food recognition aims at labeling each food region with the identity of the depicted food, and food analysis aims at determining properties of the food such as its quantity or ingredients.

We investigate the negative effects of rain streaks over the performance of a neural network for real time semantic segmentation of street scenes. This is done by synthetically augmenting the CityScapes dataset with artificial rain. We then define and train a generative adversarial network for rain removal, and quantify the benefits of its application as a pre-processing step to both rainy and “clean” images. Finally, we show that by retraining the semantic segmentation network on images processed for rain removal, it is possible to gain even more accuracy, with a model that produces stable results in all analyzed atmospheric conditions. For our experiments, we present a per-class analysis in order to provide deeper insights over the impact of rain on semantic segmentation.

Color constancy algorithms are typically evaluated with a statistical analysis of the recovery angular error and the reproduction angular error between the estimated and ground truth illuminants. Such analysis provides information about only the magnitude of the errors, and not about their chromatic properties. We propose an Angle-Retaining Chromaticity diagram (ARC) for the visual analysis of the estimated illuminants and the corresponding errors. We provide both quantitative and qualitative proof of the superiority of ARC in preserving angular distances compared to other chromaticity diagrams, making it possible to quantify the reproduction and recovery errors in terms of Euclidean distances on a plane. We present two case studies for the application of the ARC diagram in the visualization of the ground truth illuminants of color constancy datasets, and the visual analysis of error distributions of color constancy algorithms.

We sought to develop a minimally invasive intra-amniotic therapy for prenatal treatment of myelomeningocele (MMC) in an established rat model.Time-dated pregnant rats were gavage-fed retinoic acid to induce MMC. Groups received intraamniotic injections at E17.5 with alginate particles loaded with fluorescent dye, basic fibroblast growth factor (Alg-HSA-bFGF), fluorescently tagged albumin (Alginate-BSA-TR), free bFGF, blank alginate particles (Alg-Blank), or PBS. Groups were analyzed at 3 h for specific particle binding or at term (E21) to determine MMC coverage.Alginate microparticles demonstrated robust binding to the MMC defect 3 h after injection. Of those specimens analyzed at E21, 150 of 239 fetuses (62.8%) were viable. Moreover, 18 of 61 (30%) treated with Alg-HSA-bFGF showed evidence of soft tissue coverage compared to 0 of 24 noninjected (P = 0.0021), 0 of 13 PBS (P = 0.0297), and 0 of 42 free bFGF (P = P 0.0001). Scaffolds of aggregated particles associated with disordered keratinized tissue were observed covering the defect in 2 of 18 (11%) Alg-BSA-TR and 3 of 19 (16%) Alg-Blank specimens.Injection of microparticles loaded with bFGF resulted in significant soft tissue coverage of the MMC defect compared to controls. Alginate microparticles without growth factors might result in scaffold development over the fetal MMC.Basic science.N/A.

Objective The objective of this work was to investigate technical validation and usability of an innovative, technically simple, easy-to-use, and portable integrated system to assess carotid function and structure by ultrasound. Methods The studied system integrated a hardware (the Interson SP-L01 embedded ultrasound probe [Interson, Pleasanton, CA, USA]) and a software measuring the instantaneous diameter of the carotid artery in real-time from B-mode ultrasound image sequences (Carotid Studio, by Quipu Srl [Pisa, Italy]). Technical validation was evaluated by intra-operator reproducibility of two measurements acquired by an expert operator, and agreement with state-of-the-art technique (Mylab25 by Esaote SpA [Genova, Italy], Carotid Studio 4.3 by Quipu Srl) was evaluated in laboratory settings in 12 healthy volunteers; usability of the portable integrated system was investigated by administering questionnaires to users and the results were reported with scores based on a five-point scale. Results Twelve (12) healthy volunteers (five men, mean age 44.5±13.6 years, free of cardiovascular disease or risk factors), were recruited. Agreement with state-of-the-art technique was satisfactory, with no significant bias. Coefficient of variation (intra-operator reproducibility) was 3.2% (2.5% SD) for intima-media thickness, 0.9% (0.7% SD) for diameter, and 2.5% (2.2% SD) for distension. Usability questionnaires showed an overall positive judgement of the integrated system with respect to the traditional one, obtaining an average score greater than 4 (on a five-point scale). Conclusions A portable, innovative prototype to easily assess ultrasound carotid parameters of vascular ageing was successfully designed, developed, and demonstrated to be comparable with state-of-the art technique. Usability was also satisfactory.

Intraamniotic microparticle injection is a novel technique for the treatment of myelomeningocele (MMC) in which microparticles are delivered in-utero in a minimally invasive fashion to bind to and protect the exposed spinal cord. This technique could offer earlier intervention and greater access to prenatal treatment of MMC. Here we demonstrate progress on the engineering of the microparticles to promote binding to the MMC defect. We hypothesized that when the particle's surface charge was decreased and delivery concentration increased, particles would bind to the MMC defect more frequently and more specifically.Alginate microparticles underwent surface modification to alter the particle charge. Dye-loaded alginate, alginate- dextran sulfate, and alginate- chitosan were injected on e17 into the amnion of a rat model of MMC and the incidence of successful binding and specificity of particle binding to the MMC defect were calculated. Specificity of binding was described using a defect-to-skin brightness ratio based on specimen imaging. Comparisons were made with chi-square, p 0.05 marked significance.There was no difference in the incidence of successful binding at e17 with 0.6 mg/fetal kg between the three tested alginate particles. However, alginate- dextran sulfate bound most specifically to the defect (p 0.05). Alginate-dextran sulfate also demonstrated more frequent binding at higher doses than lower doses (79% at 1.2 mg/kg vs 38% at 0.6 mg/kg and 24% at 0.8 mg/kg, p 0.01 for both). Specificity was not sacrificed at higher dose injections: defect-to-skin brightness ratio of 5.4 at 1.2 mg/kg vs 1.8 at 0.6 mg/kg (p 0.05) CONCLUSION: We demonstrate that the intraamniotic injection of alginate-dextran sulfate microparticles at high concentration bind more frequently and more specifically to MMC defects than the previously tested unmodified alginate microparticles.

This article illustrates how you can assign a metabolic exercise to a tennis player, after having performed the "Sigma Test" and having acquired the subjective parameters.

We propose a new representation of distance information that is independent from any specific acquisition device, based on the size of portrayed subjects. In this alternative description, each pixel of an image is associated with the size, in real life, of what it represents. Using our proposed representation, datasets acquired with different devices can be effortlessly combined to build more powerful models, and monocular distance estimation can be performed on images acquired from devices that were never used during training. To assess the advantages of the proposed representation, we used it to train a fully convolutional neural network that predicts with pixel-precision the size of different subjects depicted in the image, as a proxy for their distance. Experimental results show that our representation, allowing the combination of heterogeneous training datasets, makes it possible for the trained network to gain better results at test time.

QUBIC (Q \& U Bolometric Interferometer for Cosmology) is an international ground-based experiment dedicated in the measurement of the polarized fluctuations of the Cosmic Microwave Background (CMB). It is based on bolometric interferometry, an original detection technique which combine the immunity to systematic effects of an interferometer with the sensitivity of low temperature incoherent detectors. QUBIC will be deployed in Argentina, at the Alto Chorrillos mountain site near San Antonio de los Cobres, in the Salta province. The QUBIC detection chain consists in 2048 NbSi Transition Edge Sensors (TESs) cooled to 350mK.The voltage-biased TESs are read out with Time Domain Multiplexing based on Superconducting QUantum Interference Devices (SQUIDs) at 1 K and a novel SiGe Application-Specific Integrated Circuit (ASIC) at 60 K allowing to reach an unprecedented multiplexing (MUX) factor equal to 128. The QUBIC experiment is currently being characterized in the lab with a reduced number of detectors before upgrading to the full instrument. I will present the last results of this characterization phase with a focus on the detectors and readout system., Conference proceedings submitted to the Journal of Low Temperature Physics for LTD18

We address the task of image saliency estimation through proper recombination of existing methods in the state of the art. We define a general scheme, which we then specialize to perform dataset-specific and image-specific recombination, based on either linear weight regression, or method selection. The advantage of this approach lies in the possibility of exploiting the different strengths of existing methods. Experiments are conducted with both deep learning and hand-crafted methods on a widely used dataset, using standard evaluation measures. The proposed recombination strategy allows us to improve upon the state of the art, by exploiting a linear combination of the saliency maps produced by existing methods. We also show that image-specific combination and selection of saliency maps is limited by the apparent lack of relevant information intrinsic in the image itself.

Introduction In peacetime, it is challenging for Army Forward Resuscitative Surgical Teams (FRST) to maintain combat readiness as trauma represents Methods Over a 3-year period, all FRSTs were surveyed at one civilian center. Results Six hundred and thirteen FRST soldiers were provided surveys and 609 responded (99.3%), including 499 (81.9%) non-physicians and 110 (18.1%) physicians/physician assistants. The non-physician group included 69% male with an average age of 34 ± 11 years and consisted of 224 AR (45%) and 275 AD (55%). Rank ranged from Private to Colonel with officers accounting for 41%. For AD vs. AR, combat experience was similar: 50% vs. 52% had ≥1 combat deployment, 52% vs. 60% peri-deployment patient load was trauma-related, and 31% vs. 32% had ≥40 patient contacts during most recent deployment (all P > .15). However, medical experience differed for AD and AR: 18% vs. 29% had >15 years of experience in practice and 4% vs. 17% spent >50% of their time treating critically injured patients (all P Conclusions This is the first study of clinical practice patterns in AD vs. AR, non-physician members of Army FRSTs. In concordance with previous studies of military surgeons, FRST non-physicians seem to be lacking clinical experience as well. To maintain readiness and to provide optimal care for our injured warriors, the entire FRST, not just individuals, should embed within civilian centers.

Purpose Contact tracing has proven successful at controlling coronavirus 2019 (COVID-19) globally, and the Center for Health Security has recommended that the United States add 100,000 contact tracers to the current workforce. Methods To address gaps in local contact tracing, health professional students partnered with their academic institution to conduct contact tracing for all COVID-19 cases diagnosed onsite, which included identifying and reaching their contacts, educating participants, and providing social resources to support effective quarantine and isolation. Results From March 24 to May 28, 536 laboratory-confirmed COVID-19 cases were contacted and reported an average of 2.6 contacts. Contacts were informed of their exposure, asked to quarantine, and monitored for the onset of symptoms. Callers reached 94% of cases and 84% of contacts. Seventy-four percent of cases reported at least one contact. Household members had higher rates of reporting symptoms (odds ratio, 1.65; 95% confidence interval, 1.19–2.28). The average test turnaround time decreased from 21.8 days for the first patients of this program to 2.3 days on the eleventh week. Conclusions This provides evidence for the untapped potential of community contact tracing to respond to regional needs, confront barriers to effective quarantine, and mitigate the spread of COVID-19.

Medical schools are creating programs to introduce careers in healthcare to students at earlier stages in their education. At the Penn State College of Medicine University Park Regional Campus, we have created and implemented a novel improvisational, case-based, small-group learning curriculum for high school students to introduce integrative thinking in basic and clinical sciences and to cultivate an early interest in health professions careers. The program also includes elements of health systems science and humanities. A multidisciplinary team of educators with both secondary and post-secondary pedagogical expertise worked together to develop a longitudinal case-based curriculum suitable for high school students. This curriculum was administered to a group of 40 students in grades 9-12. At the start of each session, students were challenged to explore personal biases, reflect upon ethical dilemmas, and to step outside their comfort zones with “centering exercises.” Through these activities, we strived to challenge the learners’ preconceptions about their colleagues, medicine, and their role within it. Students were then presented with a case as a large group. At critical junctures in the cases, students worked in small groups of 8-10 with 1-2 medical student facilitators to discuss questions and solve clinical dilemmas associated with the case. Early feedback from students and faculty advisors has been overwhelmingly positive. The authors have no conflicts of interest.

OBJECTIVE: Insufficient sleep is common among caregivers and is associated with worse health outcomes; however, the contributors to poor sleep among caregivers are unknown. We investigated the cross-sectional association between socioeconomic status (SES), psychosocial stressors and sleep among caregivers. METHODS: Caregivers (n=98) of teenagers with asthma self-reported sleep duration (hours), sleep quality (very good to very bad), education (< high school to college graduate), income ($15,000 to $50,000), and psychosocial stressors (stress, worry, social support, depressive symptoms, nightly awakenings due to caregiving). Logistic and linear regression models were performed to study the association of between SES, psychosocial stressors and sleep, adjusting for possible confounders. RESULTS: Caregivers on average were 45.5 years, female (89%) and African American (90%). Average sleep duration was 5.9 hours (standard deviation: 1.5), 72% reported short sleep (

We address the task of classifying car images at multiple levels of detail, ranging from the top-level car type, down to the specific car make, model, and year. We analyze existing datasets for car classification, and identify the CompCars as an excellent starting point for our task. We show that convolutional neural networks achieve an accuracy above 90% on the finest-level classification task. This high performance, however, is scarcely representative of real-world situations, as it is evaluated on a biased training/test split. In this work, we revisit the CompCars dataset by first defining a new training/test split, which better represents real-world scenarios by setting a more realistic baseline at 61% accuracy on the new test set. We also propagate the existing (but limited) type-level annotation to the entire dataset, and we finally provide a car-tight bounding box for each image, automatically defined through an ad hoc car detector. To evaluate this revisited dataset, we design and implement three different approaches to car classification, two of which exploit the hierarchical nature of car annotations. Our experiments show that higher-level classification in terms of car type positively impacts classification at a finer grain, now reaching 70% accuracy. The achieved performance constitutes a baseline benchmark for future research, and our enriched set of annotations is made available for public download.

INTRODUCTION Previous studies have demonstrated the importance of adherence to oral tyrosine kinase inhibitors in improving outcomes, including achieving a complete cytogenic response. Patients with chronic myeloid leukemia (CML) that had a suboptimal response were more likely to be non-adherent. Early identification and intervention based on predictors of non-adherence may lead to improved outcomes for patients in the non-trial setting. This research aimed to determine the rate of adherence and persistence to oral tyrosine kinase inhibitors (TKI) and to assess associated effect of patient characteristics using real world data from a retail pharmacy setting. METHODS This retrospective analysis of administrative pharmacy claims data included a random sample of 5000 patients who filled at least one TKI medication (bosutinib, dasatinib, imatinib, nilotinib) from national retail pharmacy chain in the study period of May 1, 2018 to April 30, 2021. Data elements included prescription fill attributes, patient-level demographics, medication adherence by therapeutic class (TKI, antidiabetics, antihypertensives and antihyperlipidemics), as well as patient health conditions and diagnoses. Patient adherence barrier data were also analyzed for a subset of patients who received select clinical interventions. This research was reviewed and approved by Advarra IRB as exempt (Pro00044844). Medication adherence was measured using the proportion of days covered (PDC) metric. For each therapeutic class, PDC was measured from first fill date for that class from May 01 2018 to April 30 2021, followed for maximum of 365 days, and calculated as the ratio of the number of days of medication available and the measurement period. A cut-off to indicate suboptimum adherence of 45 days. PDC and length of therapy and their influential factors were assessed using generalized linear models. Persistency rates were calculated descriptively and using Kaplan-Meier analyses. Associations among PDCs in TKI and common chronic medications were assessed using multivariate correlation statistics. All statistics were conducted using SAS 9.4. RESULTS The random sample of patients had a mean age of 61.7 years (median=65.6, IQR= =51, 75) and 49.9% were male. TKI use in our sample was predominantly imatinib (2,857, 57.1%) and dasatinib (1,428, 28.6%) with fewer patients on nilotinib (556, 11.1%) or bosutinib (159, 3.2%). Among those patients with adherence barrier data, the average number of barriers was 1.7. Among TKI users, 38.3% also had hypertension, 13.6% had diabetes and 15.6% had hyperlipidemia. Percent of TKI users who had common chronic conditions and who were taking corresponding therapies were 38.2%, 33.9%, and 33.3% for hypertension, hyperlipidemia, and diabetes, correspondingly. Mean TKI PDC was 0.797 (95% CL 0.789 to 0.804) with a median of 0.889. Over half of patients (55% patients) had a PDC>=.85. Mean TKI length of therapy was 18.3 months with a median of 15 months with differences by therapy. Correlation of TKI PDC to three common chronic therapy PDC were all low: 0.095 for antidiabetics; 0.032 for antihypertensives; 0.083 for antilipidemics) and not statistically significant. Age was a significant predicator of PDC, with every 10-year increase in age associated with a 2% increase in PDC. When a patient had previously stated adherence barriers, PDC was estimated to decrease by 1.7% for each barrier faced by the patient. Only a small portion (33%-38%) of TKI patients, who had a diagnosis for a common chronic condition, had a claim for the corresponding therapeutic class. CONCLUSIONS Adherence to TKI was influenced by non-modifiable risk such as age and modifiable risks such as the number of adherence barriers. Many patients on TKI who also had a common chronic condition were not taking medications for their chronic condition, noting a discordance in care. Ongoing capture of barrier data beyond specialty medications will help predict patient adherence behavior and identify targeted interventions. Disclosures No relevant conflicts of interest to declare.

We present a review of methods for automatic estimation of visual saliency: the perceptual property that makes specific elements in a scene stand out and grab the attention of the viewer. We focus on domains that are especially recent and relevant, as they make saliency estimation particularly useful and/or effective: omnidirectional images, image groups for co-saliency, and video sequences. For each domain, we perform a selection of recent methods, we highlight their commonalities and differences, and describe their unique approaches. We also report and analyze the datasets involved in the development of such methods, in order to reveal additional peculiarities of each domain, such as the representation used for the ground truth saliency information (scanpaths, saliency maps, or salient object regions). We define domain-specific evaluation measures, and provide quantitative comparisons on the basis of common datasets and evaluation criteria, highlighting the different impact of existing approaches on each domain. We conclude by synthesizing the emerging directions for research in the specialized literature, which include novel representations for omnidirectional images, inter- and intra- image saliency decomposition for co-saliency, and saliency shift for video saliency estimation.

Assisted living technologies can be of great importance for taking care of elderly people and helping them to live independently. In this work, we propose a monitoring system designed to be as unobtrusive as possible, by exploiting computer vision techniques and visual sensors such as RGB cameras. We perform a thorough analysis of existing video datasets for action recognition, and show that no single dataset can be considered adequate in terms of classes or cardinality. We subsequently curate a taxonomy of human actions, derived from different sources in the literature, and provide the scientific community with considerations about the mutual exclusivity and commonalities of said actions. This leads us to collecting and publishing an aggregated dataset, called ALMOND (Assisted Living MONitoring Dataset), which we use as the training set for a vision-based monitoring approach.We rigorously evaluate our solution in terms of recognition accuracy using different state-of-the-art architectures, eventually reaching 97% on inference of basic poses, 83% on alerting situations, and 71% on daily life actions. We also provide a general methodology to estimate the maximum allowed distance between camera and monitored subject. Finally, we integrate the defined actions and the trained model into a computer-vision-based application, specifically designed for the objective of monitoring elderly people at their homes.

QUBIC, the Q & U Bolometric Interferometer for Cosmology, is a novel ground-based instrument that aims to measure the extremely faint B-mode polarisation anisotropy of the cosmic microwave background at intermediate angular scales (multipoles of = 30 − 200). Primordial B-modes are a key prediction of Inflation as they can only be produced by gravitational waves in the very early universe. To achieve this goal, QUBIC will use bolometric interferometry, a technique that combines the sensitivity of an imager with the immunity to systematic effects of an interferometer. It will directly observe the sky through an array of back-to-back entry horns whose beams will be superimposed using a cooled quasioptical beam combiner. Images of the resulting interference fringes will be formed on two focal planes, each tiled with transition-edge sensors, cooled down to 320 mK. A dichroic filter placed between the optical combiner and the focal planes will select two frequency bands (centred at 150 GHz and 220 GHz), one frequency per focal plane. Polarization modulation will be achieved using a cold stepped half-wave plate (HWP) and polariser in front of the sky-facing horns. The full QUBIC instrument is described elsewhere1,2,3,4; in this paper we will concentrate in particular on simulations of the optical combiner (an off-axis Gregorian imager) and the feedhorn array. We model the optical performance of both the QUBIC full module and a scaled-down technological demonstrator which will be used to validate the full instrument design. Optical modelling is carried out using full vector physical optics with a combination of commercial and in-house software. In the high-frequency channel we must be careful to consider the higher-order modes that can be transmitted by the horn array. The instrument window function is used as a measure of performance and we investigate the effect of, for example, alignment and manufacturing tolerances, truncation by optical components and off-axis aberrations. We also report on laboratory tests carried on the QUBIC technological demonstrator in advance of deployment to the observing site in Argentina.

International audience; QUBIC (the Q and U Bolometric Interferometer for Cosmology) is a ground-based experiment which seeks to improve the current constraints on the amplitude of primordial gravitational waves. It exploits the unique technique, among Cosmic Microwave Background experiments, of bolometric interferometry, combining together the sensitivity of bolometric detectors with the control of systematic effects typical of interferometers. QUBIC will perform sky observations in polarization, in two frequency bands centered at 150 and 220 GHz, with two kilo-pixel focal plane arrays of NbSi Transition-Edge Sensors (TES) cooled down to 350 mK. A subset of the QUBIC instrument, the so called QUBIC Technological Demonstrator (TD), with a reduced number of detectors with respect to the full instrument, will be deployed and commissioned before the end of 2018. The voltage-biased TES are read out with Time Domain Multiplexing and an unprecedented multiplexing (MUX) factor equal to 128. This MUX factor is reached with two-stage multiplexing: a traditional one exploiting Superconducting QUantum Interference Devices (SQUIDs) at 1K and a novel SiGe Application-Specific Integrated Circuit (ASIC) at 60 K. The former provides a MUX factor of 32, while the latter provides a further 4. Each TES array is composed of 256 detectors and read out with four modules of 32 SQUIDs and two ASICs. A custom software synchronizes and manages the readout and detector operation, while the TES are sampled at 780 Hz (100kHz/128 MUX rate). In this work we present the experimental characterization of the QUBIC TES arrays and their multiplexing readout chain, including time constant, critical temperature, and noise properties.

QUBIC, the Q & U Bolometric Interferometer for Cosmology, is a novel ground-based instrument that has been designed to measure the extremely faint B-mode polarisation anisotropy of the cosmic microwave background at intermediate angular scales (multipoles of 𝑙 = 30 − 200). Primordial B-modes are a key prediction of Inflation as they can only be produced by gravitational waves in the very early universe. To achieve this goal, QUBIC will use bolometric interferometry, a technique that combines the sensitivity of an imager with the systematic error control of an interferometer. It will directly observe the sky through an array of 400 back-to-back entry horns whose signals will be superimposed using a quasi-optical beam combiner. The resulting interference fringes will be imaged at 150 and 220 GHz on two focal planes, each tiled with NbSi Transition Edge Sensors, cooled to 320 mK and read out with time-domain multiplexing. A dichroic filter placed between the optical combiner and the focal planes will select the two frequency bands. A very large receiver cryostat will cool the optical and detector stages to 40 K, 4 K, 1 K and 320 mK using two pulse tube coolers, a novel 4He sorption cooler and a double-stage 3He/4He sorption cooler. Polarisation modulation and selection will be achieved using a cold stepped half-wave plate (HWP) and polariser, respectively, in front of the sky-facing horns. A key feature of QUBIC’s ability to control systematic effects is its ‘self-calibration’ mode where fringe patterns from individual equivalent baselines can be compared. When observing, however, all the horns will be open simultaneously and we will recover a synthetic image of the sky in the I, Q and U Stokes’ parameters. The synthesised beam pattern has a central peak of approximately 0.5 degrees in width, with secondary peaks further out that are damped by the 13-degree primary beam of the horns. This is Module 1 of QUBIC which will be installed in Argentina, near the city of San Antonio de los Cobres, at the Alto Chorrillos site (4869 m a.s.l.), Salta Province. Simulations have shown that this first module could constrain the tensor-to-scalar ratio down to σ(r) = 0.01 after a two-year survey. We aim to add further modules in the future to increase the angular sensitivity and resolution of the instrument. The QUBIC project is proceeding through a sequence of steps. After an initial successful characterisation of the detection chain, a technological demonstrator is being assembled to validate the full instrument design and to test it electrically, thermally and optically. The technical demonstrator is a scaled-down version of Module 1 in terms of the number of detectors, input horns and pulse tubes and a reduction in the diameter of the combiner mirrors and filters, but is otherwise similar. The demonstrator will be upgraded to the full module in 2019. In this paper we give an overview of the QUBIC project and instrument.

In this paper we present a deep learning method to estimate the illuminant of an image. Our model is not trained with illuminant annotations, but with the objective of improving performance on an auxiliary task such as object recognition. To the best of our knowledge, this is the first example of a deep learning architecture for illuminant estimation that is trained without ground truth illuminants. We evaluate our solution on standard datasets for color constancy, and compare it with state of the art methods. Our proposal is shown to outperform most deep learning methods in a cross-dataset evaluation setup, and to present competitive results in a comparison with parametric solutions., Accepted at ICIP 2018

We focus on saliency estimation in digital images. We describe why it is important to adopt a data-driven model for such an illposed problem, allowing for a universal concept of “saliency” to naturally emerge from data that are typically annotated with drastically heterogeneous criteria. Our learning-based method also involves an explicit analysis of the input at multiple scales, in order to take into account images of different resolutions, depicting subjects of different sizes. Furthermore, despite training our model on binary ground truths only, we are able to output a continuous-valued confidence map, which represents the probability of each image pixel being salient. Every contribution of our method for saliency estimation is singularly tested according to a standard evaluation benchmark, and our final proposal proves to be very effective in a comparison with the state-of-the-art.

We present a state of the art method for vegetable and fruit recognition based on convolutional neural networks. We developed our solution around the concept of a smart kitchen/refrigerator equipped with an on-board camera. With this objective in mind, we adopted a dataset that was specifically collected and annotated according to the eating characteristics of the portrayed items. We performed two types of experiment: we first trained and evaluated different state-of-the-art neural architectures on the task of vegetable and fruit recognition. Secondly, we designed and tested a solution that exploits the hierarchical nature of such classes to further improve the final performance of our system. Experimental results demonstrate the quantitative superiority of the proposed solution compared to existing approaches.

We propose a novel modular CNN architecture that provides semantic segmentation and understanding of outdoor street environment images. Our solution processes a 512x1024 resolution image on a single Titan Xp GPU at 37.4 FPS attaining 70.4% IoU on the Cityscapes test dataset.

International audience; Remnant radiation from the early universe, known as the Cosmic Microwave Background (CMB), has been redshifted and cooled, and today has a blackbody spectrum peaking at millimetre wavelengths. The QUBIC (Q&U Bolometric Interferometer for Cosmology) instrument is designed to map the very faint polaristion structure in the CMB. QUBIC is based on the novel concept of bolometric interferometry in conjunction with synthetic imaging. It will have a large array of input feedhorns, which creates a large number of interferometric baselines. The beam from each feedhorn is passed through an optical combiner, with an off-axis compensated Gregorian design, to allow the generation of the synthetic image. The optical-combiner will operate in two frequency bands (150 and 220 GHz with 25% and 18.2 % bandwidth respectively) while cryogenically cooled TES bolometers provide the sensitivity required at the image plane. The QUBIC Technical Demonstrator (TD), a proof of technology instrument that contains 64 input feed-horns, is currently being built and will be installed in the Alto Chorrillos region of Argentina. The plan is then for the full QUBIC instrument (400 feed-horns) to be deployed in Argentina and obtain cosmologically significant results. In this paper we will examine the output of the manufactered feed-horns in comparison to the nominal design. We will show the results of optical modelling that has been performed in anticipation of alignment and calibration of the TD in Paris, in particular testing the validity of real laboratory environments. We show the output of large calibrator sources (50 ° full width haf max Gaussian beams) and the importance of accurate mirror definitions when modelling large beams. Finally we describe the tolerance on errors of the position and orientation of mirrors in the optical combiner.

International audience; QUBIC, the QU Bolometric Interferometer for Cosmology, is a novel forthcoming instrument to measure the B-mode polarization anisotropy of the Cosmic Microwave Background. The detection of the B-mode signal will be extremely challenging; QUBIC has been designed to address this with a novel approach, namely bolometric interferometry. The receiver cryostat is exceptionally large and cools complex optical and detector stages to 40 K, 4 K, 1 K and 350 mK using two pulse tube coolers, a novel 4He sorption cooler and a double-stage 3He/4He sorption cooler. We discuss the thermal and mechanical design of the cryostat, modelling and thermal analysis, and laboratory cryogenic testing.

In the Province of Ontario, Canada, aggressive policy promoting wind energy development has led to both rapid development and intense stakeholder conflicts. Focusing on developers, key stakeholders largely hidden in the extant research literature and the perspectives of other stakeholders drawn from secondary sources, this paper presents original primary research to help fill this knowledge gap. Based on semi-structured interviews with established and active developers in the Province, we find that feed-in tariffs have arguably been the strongest driver of developers successfully getting turbines up and running. Yet, legislative and policy attempts to reduce delays and smooth the development process have often complicated the development process. Developers recognise and often agree with community viewpoints that the process as framed by Ontario’s policy environment forestalls cooperative development, particularly with respect to community engagement. While developers are supportive of better community engagement, they feel constrained by policy-related barriers. Findings from the study show that communities will only be engaged in projects to the full extent possible if developers take the initiative to transcend regulatory requirements for public engagement. The study concludes with useful lessons for jurisdictions transitioning to low or zero emissions energy technologies. Specifically, it supports recommendations for alternative policy approaches including consideration of policy specificity around economic benefit destitution, and community engagement and ownership of projects.

BACKGROUND: Portable noninvasive Heart Rate Complexity (HRC) and Compensatory Reserve Measurement (CRM) monitors have been developed to triage supine combat casualties. Neither monitor has been tested in upright individuals during physical exercise. This study tests the hypothesis that exercise evokes proportional changes in HRC and CRM.METHODS: Two instruments monitored volunteers (9 civilian and 11 soldiers) from the Army Trauma Training Department (ATTD) before, during, and following physical exercise. One recorded heart rate (HR, bpm), cardiac output (CO, L · min−1), heart rate variability (HRV, root mean square of successive differences, ms), and HRC (Sample Entropy, unitless). The other recorded HR, pulse oximetry (Spo2, %), and CRM (%).RESULTS: Baseline HR, CO, HRV, HRC, and CRM averaged 72 ± 1 bpm, 5.6 ± 1.2 L · min−1, 48 ± 24 ms, 1.9 ± 0.5, and 85 ± 10% in seated individuals. Exercise evoked peak HR and CO at > 200% of baseline, while HRC and CRM were simultaneously decreased to minimums that were ≤ 50% of baseline (all P < 0.001). HRV changes were variable and unreliable. Spo2 remained consistently above 95%. During a 60 min recovery, HR and CRM returned to baseline on parallel tracks (t1/2=11 ± 8 and 18 ± 14 min), whereas HRC recovery was slower than either CRM or HR (t1/2=40 ± 18 min, both P < 0.05).DISCUSSION: Exercise evoked qualitatively similar changes in CRM and HRC. CRM recovered incrementally faster than HRC, suggesting that vasodilation, muscle pump, and respiration compensate faster than cardiac autonomic control in young, healthy volunteers. Both HRC and CRM appear to provide reliable, objective, and noninvasive metrics of human performance in upright exercising individuals.Mulder MB, Eidelson SA, Buzzelli MD, Gross KR, Batchinsky AI, Convertino VA, Schulman CI, Namias N, Proctor KG. Exercise-induced changes in compensatory reserve and heart rate complexity. Aerosp Med Hum Perform. 2019; 90(12):1009–1015.

The recruitment setting plays a key role in the evaluation of behavioral interventions. We evaluated a behavioral intervention for urban adolescents with asthma in three randomized trials conducted separately in three different settings over the course of 8 years. We hypothesized that characteristics of trial participants recruited from the ED and clinic settings would be significantly different from that of youth participating in the school-based trials. The intervention evaluated was Puff City, a web-based program that uses tailoring to improve asthma management behaviors.The present analysis includes youth aged 13-19 years who reported a physician diagnosis of asthma and symptoms at trial baseline. In the three trials, all participants were randomized post-baseline to a web-based, tailored intervention (treatment) or generic web-based asthma education (control).Compared to school-based trial participants, ED participants had significantly more acute-care visits for asthma (Compared to participants in the school-based trials, participants recruited from ED and clinic settings differed significantly in terms of healthcare use, as well as psychosocial and sociodemographic factors. These factors can inform intervention content, and may impact external validity of behavioral interventions for asthma.

In this paper we address the task of salient object detection without requiring an explicit object class recognition. To this end, we propose a solution that exploits intermediate activations of a Fully Convolutional Neural Network previously trained for the recognition of 1,000 object classes, in order to gather generic object information at different levels of resolution. This is done by using both convolution and convolution-transpose layers, and combining their activations to generate a pixel-level salient object segmentation. Experiments are conducted on a standard benchmark that involves seven heterogeneous datasets. On average our solution outperforms the state of the art according to multiple evaluation measures.

In this paper we propose a method for logo recognition using deep learning. Our recognition pipeline is composed of a logo region proposal followed by a Convolutional Neural Network (CNN) specifically trained for logo classification, even if they are not precisely localized. Experiments are carried out on the FlickrLogos-32 database, and we evaluate the effect on recognition performance of synthetic versus real data augmentation, and image pre-processing. Moreover, we systematically investigate the benefits of different training choices such as class-balancing, sample-weighting and explicit modeling the background class (i.e. no-logo regions). Experimental results confirm the feasibility of the proposed method, that outperforms the methods in the state of the art., Preprint accepted in Neurocomputing

Transitioning to renewable energy has become a global phenomenon. Attempts to explain development patterns are often fragmented, focusing on specific rationale for (un)successful deployment or the perspective of specific stakeholders. We recognize the importance for holistic insights for successful and effective environmental policy-making and planning in the context of renewable energy development. Using wind energy as an example, this paper provides methodological toolkits for utilizing content analysis (CA) to gain more holistic insights on renewable energy deployment and policy outcomes. Building on existing literature, past and ongoing research, three variations of CA are presented for understanding stakeholder conflicts surrounding wind energy development as well as public perceptions and responses to the technology across space and over time. Although newspapers are used as an example, these methodological protocols could be applied to diverse forms of data such as interviews, policy docume...

CMB B-mode experiments are required to control systematic effects with an unprecedented level of accuracy. Polarization modulation by a half wave plate (HWP) is a powerful technique able to mitigate a large number of the instrumental systematics. Our goal is to optimize the polarization modulation strategy of the upcoming LSPE-SWIPE balloon-borne experiment, devoted to the accurate measurement of CMB polarization at large angular scales. We depart from the nominal LSPE-SWIPE modulation strategy (HWP stepped every 60 s with a telescope scanning at around 12 deg/s) and perform a thorough investigation of a wide range of possible HWP schemes (either in stepped or continuously spinning mode and at different azimuth telescope scan-speeds) in the frequency, map and angular power spectrum domain. In addition, we probe the effect of high-pass and band-pass filters of the data stream and explore the HWP response in the minimal case of one detector for one operation day (critical for the single-detector calibration process). We finally test the modulation performance against typical HWP-induced systematics. Our analysis shows that some stepped HWP schemes, either slowly rotating or combined with slow telescope modulations, represent poor choices. Moreover, our results point out that the nominal configuration may not be the most convenient choice. While a large class of spinning designs provides comparable results in terms of pixel angle coverage, map-making residuals and BB power spectrum standard deviations with respect to the nominal strategy, we find that some specific configurations (e.g., a rapidly spinning HWP with a slow gondola modulation) allow a more efficient polarization recovery in more general real-case situations. Although our simulations are specific to the LSPE-SWIPE mission, the general outcomes of our analysis can be easily generalized to other CMB polarization experiments., 11 pages, 9 figures, accepted for publication in A&A

The objective of this study was to review environmental equity research on outdoor air pollution and, specifically, methods and tools used in research, published in English, with the aim of recommending the best methods and analytic tools. English language publications from 2000 to 2012 were identified in Google Scholar, Ovid MEDLINE, and PubMed. Research methodologies and results were reviewed and potential deficiencies and knowledge gaps identified. The publications show that exposure to outdoor air pollution differs by social factors, but findings are inconsistent in Canada. In terms of study designs, most were small and ecological and therefore prone to the ecological fallacy. Newer tools such as geographic information systems, modeling, and biomarkers offer improved precision in exposure measurement. Higher-quality research using large, individual-based samples and more precise analytic tools are needed to provide better evidence for policy-making to reduce environmental inequities.

The search for the B-mode polarization of Cosmic Microwave Background (CMB) is the new frontier of observational Cosmology. A B-mode detection would give an ultimate confirmation to the existence of a primordial Gravitational Wave (GW) background as predicted in the inflationary scenario. Several experiments have been designed or planned to observe B-modes. In this work we focus on the forthcoming Large Scale Polarization Explorer (LSPE) experiment, that will be devoted to the accurate measurement of CMB polarization at large angular scales. LSPE consists of a balloon-borne bolometric instrument, the Short Wavelength Instrument for the Polarization Explorer (SWIPE), and a ground-based coherent polarimeter array, the STRatospheric Italian Polarimeter (STRIP). SWIPE will employ a rotating Half Wave Plate (HWP) polarization modulator to mitigate the systematic effects due to instrumental non-idealities. We present here preliminary forecasts aimed at optimizing the HWP configuration., 6 pages, 4 figures, proceedings of the 7th Young Researcher Meeting, Torino, Oct 24th-26th 2016