Unsupervised Action Proposals Using Support Vector Classifiers for Online Video Processing.

In this work, we introduce an intelligent video sensor for the problem of Action Proposals (AP). AP consists of localizing temporal segments in untrimmed videos that are likely to contain actions. Solving this problem can accelerate several video action understanding tasks, such as detection, retrieval, or indexing. All previous AP approaches are supervised and offline, i.e. they need both the temporal annotations of the datasets during training and access to the whole video to effectively cast the proposals. We propose here a new approach which, unlike the rest of the state-of-the-art models, is unsupervised. This implies that we do not allow it to see any labeled data during learning nor to work with any pre-trained feature on the used dataset. Moreover, our approach also operates in an online manner, which can be beneficial for many real-world applications where the video has to be processed as soon as it arrives at the sensor, e.g., robotics or video monitoring. The core of our method is based on a Support Vector Classifier (SVC) module which produces candidate segments for AP by distinguishing between sets of contiguous video frames. We further propose a mechanism to refine and filter those candidate segments. This filter optimizes a learning-to-rank formulation over the dynamics of the segments. An extensive experimental evaluation is conducted on Thumos'14 and ActivityNet datasets, and, to the best of our knowledge, this work supposes the first unsupervised approach on these main AP benchmarks. Finally, we also provide a thorough comparison to the current state-of-the-art supervised AP approaches. We achieve 41% and 59% of the performance of the best-supervised model on ActivityNet and Thumos'14, respectively, confirming our unsupervised solution as a correct option to tackle the AP problem. The code to reproduce all our results will be publicly released upon acceptance of the paper.

Boosting Multi-Vehicle Tracking with a Joint Object Detection and Viewpoint Estimation Sensor.

In this work, we address the problem of multi-vehicle detection and tracking for traffic monitoring applications. We preset a novel intelligent visual sensor for tracking-by-detection with simultaneous pose estimation. Essentially, we adapt an Extended Kalman Filter (EKF) to work not only with the detections of the vehicles but also with their estimated coarse viewpoints, directly obtained with the vision sensor. We show that enhancing the tracking with observations of the vehicle pose, results in a better estimation of the vehicles trajectories. For the simultaneous object detection and viewpoint estimation task, we present and evaluate two independent solutions. One is based on a fast GPU implementation of a Histogram of Oriented Gradients (HOG) detector with Support Vector Machines (SVMs). For the second, we adequately modify and train the Faster R-CNN deep learning model, in order to recover from it not only the object localization but also an estimation of its pose. Finally, we publicly release a challenging dataset, the GRAM Road Traffic Monitoring (GRAM-RTM), which has been especially designed for evaluating multi-vehicle tracking approaches within the context of traffic monitoring applications. It comprises more than 700 unique vehicles annotated across more than 40.300 frames of three videos. We expect the GRAM-RTM becomes a benchmark in vehicle detection and tracking, providing the computer vision and intelligent transportation systems communities with a standard set of images, annotations and evaluation procedures for multi-vehicle tracking. We present a thorough experimental evaluation of our approaches with the GRAM-RTM, which will be useful for establishing further comparisons. The results obtained confirm that the simultaneous integration of vehicle localizations and pose estimations as observations in an EKF, improves the tracking results.

A "salt and pepper" noise reduction scheme for digital images based on Support Vector Machines classification and regression.

We present a new impulse noise removal technique based on Support Vector Machines (SVM). Both classification and regression were used to reduce the "salt and pepper" noise found in digital images. Classification enables identification of noisy pixels, while regression provides a means to determine reconstruction values. The training vectors necessary for the SVM were generated synthetically in order to maintain control over quality and complexity. A modified median filter based on a previous noise detection stage and a regression-based filter are presented and compared to other well-known state-of-the-art noise reduction algorithms. The results show that the filters proposed achieved good results, outperforming other state-of-the-art algorithms for low and medium noise ratios, and were comparable for very highly corrupted images.

Assessment and counseling to get the best efficiency and effectiveness of the assistive technology (MATCH): Study protocol.

AIMS: To determine the psychosocial impact of assistive technology(AT) based on robotics and artificial intelligence in the life of people with disabilities. BACKGROUND: The best match between any person with disabilities and its AT only can be gotten through a complete assessment and monitoring of his/her needs, abilities, priorities, difficulties and limitations. Without this analysis, it's possible that the device won't meet the individual's expectations. Therefore, it is important that any project focused on the development of innovating AT for people with disabilities includes the perspective of outcome measures as an important phase of the research. In this sense, the integration of the assessment, implementation process and outcome measures is crucial to guarantee the transferability for the project findings and to get the perspective from the final user. METHODS: Pilot study, with prospective, longitudinal and analytical cohort. The study lasts from July 2020 until April 2023. The sample is formed by people with disabilities, ages from 2-21, that will participate from the first stage of the process (initial assessment of their abilities and needs) to the final application of outcome measures instruments (with a complete implication during the test of technology). DISCUSSION: Only with the active participation of the person is possible to carry out a user-centered approach. This fact will allow us to define and generate technological solutions that really adjust to the expectations, needs and priorities of the people with disabilities, avoiding the AT from being abandoned, with the consequent health and social spending. TRIAL REGISTRATION: Clinical Trials ID: NCT04723784; https://clinicaltrials.gov/.

A Low-Cost Assistive Robot for Children with Neurodevelopmental Disorders to Aid in Daily Living Activities.

In this paper, we present a new low-cost robotic platform that has been explicitly developed to increase children with neurodevelopmental disorders' involvement in the environment during everyday living activities. In order to support the children and youth with both the sequencing and learning of everyday living tasks, our robotic platform incorporates a sophisticated online action detection module that is capable of monitoring the acts performed by users. We explain all the technical details that allow many applications to be introduced to support individuals with functional diversity. We present this work as a proof of concept, which will enable an assessment of the impact that the developed technology may have on the collective of children and youth with neurodevelopmental disorders in the near future.

Octakis(3-azidopropyl)octasilsesquioxane: a versatile nanobuilding block for the efficient preparation of highly functionalized cube-octameric polyhedral oligosilsesquioxane frameworks through click assembly.

A one-step synthesis of octakis(3-azidopropyl)octasilsesquioxane from commercially available octakis(3-aminopropyl)octasilsesquioxane has been developed through a highly efficient diazo-transfer reaction under very mild conditions. Nonaflyl azide is shown to be a safer, cheaper, and more efficient reagent for this transformation than the better known and generally used diazo-transfer reagent triflyl azide. Octakis(3-azidopropyl)octasilsesquioxane is an excellent nanobuilding block that can be readily octafunctionalized with a range of terminal alkynes by copper(I)-catalyzed 1,3-dipolar azide-alkyne cycloaddition to provide new functional nanocages, maintaining a perfect 3D cubic symmetry. The mildness, simplicity, and efficiency of this approach have been demonstrated in the preparation of a glyco-polyhedral oligosilsesquioxane (POSS) conjugate and a BODIPY-POSS cluster (BODIPY = boron dipyrromethene).

Conventional unidirectional laser action enhanced by dye confined in nanoparticle scatters.

The synthesis, structural characterization, and lasing properties of new dye-sensitized organic scattering gain medium based on Rhodamine 6G (Rh6G) confined in polymeric nanoparticles are reported. We have demonstrated coherent laser action from amplifying random media using dye confined into polymeric nanoparticles as scattering centers and gain media. Lasing efficiency and photostability were significantly enhanced by nonresonant feedback of the emission by multiple scattering.

A comparative study of platelet-rich plasma, hydroxyapatite, demineralized bone matrix and autologous bone to promote bone regeneration after mandibular impacted third molar extraction.

OBJECTIVES: 1) to compare mandibular bone regeneration by applying autologous bone, platelet-rich plasma and two biomaterials (synthetic calcium hydroxyapatite, and demineralized bone matrix), and thus establish the potential benefits of these biomaterials in the regeneration of postextraction alveolar bone, 2) to identify which of them accelerates more bone regeneration and 3) to determine whether there are differences in the postoperative period (pain, swelling, trismus, infection) depending on the material used. STUDY DESIGN: It consists in a prospective, controlled (with a split- mouth design) and double blinded study. We use as a model an easily reproducible non-critical bone defect: the defect that remains after extraction of mandibular impacted third molar. The study design is based on the extraction of two mandibular impacted third molars in a patient during the same surgical procedure by the same surgeon. We assessed postoperative clinical data, and short, medium and long term neoformation of alveolar bone after extraction. We compared the two sockets (right and left), which had been grafted in a different way with the various elements mentioned above. In addition, we compared the postoperative inflammatory symptoms between groups. RESULTS: The highest acceleration in bone formation was observed in groups in which we used autologous bone and demineralized bone matrix. There were no statistically significant differences between groups regarding pain, swelling, trismus and infection throughout the postoperative period. CONCLUSIONS: According to the results of our study, autologous bone persists as the gold standard material for bone regeneration. Among the assessed biomaterials, demineralized bone matrix has yielded the best results obtained. No significant differences in the postoperative (pain, swelling, trismus and infectious events) were observed, depending on the type of material used as a graft.

Efficient red-edge materials photosensitized by Rhodamine 640.

We report on tunable, highly efficient and photostable solid-state dye laser emitting around 640 nm based on Rhodamine 640 incorporated into homopolymers, linear and cross-linked copolymers, and silicon-modified organic matrices. The effect on the lasing properties of both dye concentration and environmental conditions was analyzed. Under transversal pumping at 532 nm with 5.5 mJ/pulse, high-lasing efficiencies of up to 42% were recorded. The laser operation was highly stable with a drop in the laser output of approximately 20% after 100 000 pump pulses at the same position of the sample at 10 Hz repetition rate. To the best of our knowledge, these results are the topmost achieved to date for organic, inorganic, and hybrid materials doped with rhodamine 640. When the samples were incorporated into a grazing-incidence grating oscillator, narrow-line-width operation with tunning ranges of up to 40 nm was obtained.

Learning to Exploit the Prior Network Knowledge for Weakly-Supervised Semantic Segmentation.

Training a Convolutional Neural Network (CNN) for semantic segmentation typically requires to collect a large amount of accurate pixel-level annotations, a hard and expensive task. In contrast, simple image tags are easier to gather. With this paper we introduce a novel weakly-supervised semantic segmentation model able to learn from image labels, and just image labels. Our model uses the prior knowledge of a network trained for image recognition, employing these image annotations as an attention mechanism to identify semantic regions in the images. We then present a methodology that builds accurate class-specific segmentation masks from these regions, where neither external objectness nor saliency algorithms are required. We describe how to incorporate this mask generation strategy into a fully end-to-end trainable process where the network jointly learns to classify and segment images. Our experiments on PASCAL VOC 2012 dataset show that exploiting these generated class-specific masks in conjunction with our novel end-to-end learning process outperforms several recent weakly-supervised semantic segmentation methods that use image tags only, and even some models that leverage additional supervision or training data.

Synthesis, photophysical properties, and laser behavior of 3-amino and 3-acetamido BODIPY dyes.

The asymmetrically substituted BODIPY dyes 9a and 9b have been synthesized through a key redox step involving the alpha-nitroso derivative of the starting pyrrol. Both dyes emit fluorescence with quantum yields of ca. 0.7, but only 8b behaves as a good laser dye, with an efficiency of 48% in ethanol solution.

5-Fluorouracil plasma levels and biodegradation of subcutaneously injected drug-loaded microspheres prepared by spray-drying poly(D,L-lactide) and poly(D,L-lactide-co-glycolide) polymers.

Microspheres (MS) of 5-fluorouracil-loaded poly(D,L-lactide) (PLA), poly(D,L-lactide-co-glycolide) 75/25 (PLGA 75/25) and poly(D,L-lactide-co-glycolide) 50/50 (PLGA 50/50) prepared by the spray-drying technique were subcutaneously injected in the back of Wistar rats in order to evaluate the 5-fluorouracil (5-FU) release and the biodegradation characteristics. Determination of plasma 5-FU concentration by HPLC with analysis of data using a non-compartmental model showed drug in plasma between 9 and 14 days after administration of drug-loaded PLGA 50/50 or PLA and PLGA 75/25 microspheres, respectively, with a maximum drug concentration of 2.4+/-0.2microg/mL at 24h (5-FU-loaded PLGA 50/50 MS), 2.5+/-0.1microg/mL at 48h (5-FU-loaded PLGA 75/25 MS), and 2.3+/-0.1microg/mL at 24h (5-FU-loaded PLA MS). Pharmacokinetically, a significant increase of AUC (up to 50 times) and MRT (up to 196 times) of 5-FU with regard to the administration of the drug in solution was observed. Scanning electron microscopy and histological studies indicated that a small fibrous capsule was observed around the microspheres in the site of injection. One month after the injection of PLGA 50/50 MS and 2 months after the injection of PLGA 75/25 and PLA MS, masses of polymers, instead of single microspheres, were observed. Close to them, macrophagic cells were present, and blood vessels were observed in the connective tissue. Total absence of fibrous capsule and injected microspheres was observed after 2 (for PLGA 50/50 MS) or 3 (PLGA 75/25 and PLA MS) months.

Polymerization kinetics and mechanical characterization of new formulations of light-cured dental sealants.

The aim of the study was to develop and evaluate the in vitro performance of a new and simplified formulation of photocuring resin to be used as dental sealant. Two experimental dental sealants (CYTED1 and CYTED2) were formulated and their kinetic of polymerisation and physico-chemical properties were studied and compared to those of two commercially available sealants (Helioseal, Delton-FS). Rates of photoinitiated polymerisation (Rp), as well as the conversions and the quantum yields of polymerisation (phi(p)) were calculated. Flexural strength, Young's modulus, microhardness, microleakage, water sorption, and solubility were also tested. ANOVA, Student-Newman-Keuls, Pearson correlation and Kruskal-Wallis tests were used (p < 0.05). The highest Rp and phi(p) were obtained for the sealant CYTED2, Rp and phi(p) were similar for CYTEDl and Helioseal, and the lowest for Delton. Water sorption values were similar for Helioseal and CYTED2 being higher for CYTED1 and lower for Delton. No differences were found for solubility and microleakage values. Mechanical properties were better for Delton and no differences were found within the rest of the sealants. At short irradiation times (30 s), the maximum effectiveness of the photoinitiating system was obtained by the experimental CYTED2.

Degradation behaviour of microspheres prepared by spray-drying poly(D,L-lactide) and poly(D,L-lactide-co-glycolide) polymers.

Polymeric microsphere degradation must be taken into account in the design of drug delivery systems to be injected in in vivo systems, thus a prior analysis of in vitro degradation behaviour of microspheres appears to be necessary. In this study degradation characteristics of poly(lactide-co-glycolide) (PLGA) and poly(D,L-lactide) (PLA) microspheres prepared by the spray-drying technique have been examined. It was found that a slow decrease in molecular weight took place during the first stage of degradation, and the value of the rate constant decreased with the increase of the percentage of lactic acid of the polymer in a linear way. Thus, the period of time of this first stage decreased with the increase of content of glycolidyl units of the polymer, and it was the unique stage observed in PLA microspheres after 5 months of study. During this period of time, significant mass loss was not observed in the microspheres. The second stage of degradation of PLGA microspheres showed a larger rate constant, whose value increased with the content of glycolidyl units of the polymer. Mass loss was observed from number-average molecular weight about 6000. A sharp decrease of glass transition temperature (T(g)) was observed coinciding with the start of mass loss. This fact was accompanied by a physical change of the samples, fusion of microspheres to form large particles, which also fusion to form a unique mass of polymer; moment from that the degradation process was quicker.

Methacrylate-tethered analogs of the laser dye PM567--synthesis, copolymerization with methyl methacrylate and photostability of the copolymers.

The synthesis and characterization of new analogs of the laser dye PM567 (4,4-difluoro-1,3,5,7,8-pentamethyl-2,6-diethyl-4-bora-3a,4a-diaza-s-indacene) with the 8-position substituted by a linear chain with n methylenes (n = 1, 3, 5, 10 or 15) tethered with an acetoxy or methacryloyloxy group (PnAc and PnMA, respectively) is described. The monomeric dyes PnMA have been successfully copolymerized with methyl methacrylate (MMA), yielding linear copolymers of high optical quality where the covalently bonded chromophore is separated from the polymeric main chain by a spacer of variable length. The photostability of the solid polymeric materials under UV (310 nm) irradiation (method ASTM G 53-77) has been compared with those of the model dyes PnAc and PM567 as solid solutions in poly-MMA (PnAc-PMMA and PM567-PMMA, respectively). In all the cases, the chromophore bound to the polymer is more photostable than that simply dissolved in the same polymer, with photodegradation quantum yields in the range 2.3 x 10(-5) to 4.8 x 10(-5), which was interpreted as due to additional modes for the dissipation of the absorbed energy along the polymeric chain. In both polymer solutions and copolymers, the length of the polymethylene chain has low or null influence on the photostability. In ethanol solution, PnAc model dyes with polymethylene chains with three or more methylene groups show about the same photostability; this is of an order of magnitude higher than that of the parent dyes P1Ac and PM567 in the same solvent.

In vitro and in vivo laser treatments of tattoos: high efficiency and low fluences.

OBJECTIVE: To analyze the absorption of tattoo inks related to their in vivo and in vitro behavior under laser irradiation to improve laser-assisted tattoo removal. DESIGN: The absorption of 21 tattoo inks in a wavelength range from 300 to 800 nm was characterized by reflection spectroscopy from samples consisting of inks mixed in gelatin. Tattoo inks were removed in vitro using pulsed laser radiation with different variables, and morphologic analysis of the irradiated areas was performed. SETTING: An interdisciplinary laser laboratory with a common industrial project with the Spanish company Milesman S.A. PARTICIPANT: One person was voluntarily tattooed with 2 of the studied inks. MAIN OUTCOME MEASURES: (1) First approach to the optimum dose for pigment removal in in vitro models. (2) Correlation between the in vitro and in vivo situations at the optimum dose. RESULTS: Reflection spectroscopy facilitated selection of the most adequate laser wavelengths for tattoo removal. Red, orange, and rose inks were successfully lightened at 532 nm with 0.6 J/cm2; brown at 1064 nm with 0.3 J/cm2; yellow and green at 448 nm with 2.6 J/cm2; and blue at 600 nm with 0.9 J/cm2. Similar colors in in vitro and in vivo tattoos responded with the same efficiency to the laser variables. CONCLUSIONS: High efficiency is reached in the removal of in vivo tattoos by using an irradiation wavelength at which the percentage of reflection from the pigment is minimal. Under this condition, laser pulses can be used with a low fluence, minimizing adverse effects and clinical time.

8-PropargylaminoBODIPY: unprecedented blue-emitting pyrromethene dye. Synthesis, photophysics and laser properties.

Highly emitting 8-propargylaminoBODIPY (8-PAB) 2 was prepared in 94% yield. Unlike any other BODIPY structure hitherto described in the literature, 2 displays efficient emission in the blue region of the visible spectrum with a fluorescence quantum yield up to 0.94 and high laser efficiency (35%) at 483 nm.

Metastatic renal cell carcinoma presenting as a parotid tumor.

Metastases of malignancies to the parotid region are relatively infrequent (21%-42% of all malignant tumors 1 ), but metastases of infraclavicular origin are infrequent (0.16%-4% 1,2 ). From 1986 to 1998, only 17 cases of parotid metastases of renal clear cell carcinoma were documented in the literature reviewed. In this paper we report a case of a patient with a parotid tumor which was the first manifestation of distant disease. It was only once the tumor was surgically removed that the histopathology allowed the diagnosis of the primary tumor; namely a renal clear cell adenocarcinoma.

Complex oncologic reconstruction of a mandibular and floor of mouth defect with a fibula free flap in an achondroplastic patient.

The fibular free flap is seen as one of the foremost technical options in mandibular reconstruction, especially in those defects where long bone is required. Cases with squamous-cell carcinoma of the floor of the mouth with mandibular spread and subsequent segmentary mandibular removal are the cornerstone examples. A case of squamous-cell carcinoma of the whole floor of the mouth with mandibular invasion is reported. Radical resection of the floor of the mouth and bilateral mandibular horizontal ramus was performed, with a bony defect extending from angle to angle. The patient revealed an achondroplastic condition, with remarkable dwarfism and long-bone morphological alterations, that minimized the potential fibular length to transfer. A microsurgical reconstruction with an osteocutaneous fibular free flap was undertaken. The flap design was technically compromised by the forward bowing of the fibula and the ossification of the interosseous membrane. Specific intraoperative strategies for dealing with anatomic variations are discussed. The fibular free flap is an excellent technique for mandibular reconstruction. Morphological deviations can modify the design of the flap. Achondroplasia is not a deterrent in successful use of the free fibula flap for reconstruction of the head and neck in adequately selected cases.

Photothermal analysis of polymeric dye laser materials excited at different pump rates.

The photothermal properties and heat diffusion of polymeric lasers, made up from solutions of Rhodamine 6G in solid matrices of poly(2-hydroxyethyl methacrylate) with different amounts of the cross-linking monomer ethylene glycol dimethacrylate and copolymers of 2-hydroxyethyl methacrylate and methyl methacrylate have been studied through photothermal deflection spectroscopy. The heat load that is due to the pumping process was quantified as a function of the pump excitation repetition frequency (0.25-10 Hz), determining the time-dependent temperature changes at different locations within the laser matrix. A theoretical model, which reproduces these changes with high accuracy, was developed on the basis of the heat-diffusion equation of optically dense fluids. The observed thermal effects became important for impairing the laser stability at pump repetition frequencies higher than 1 Hz. In addition, the irreversible optical changes produced in the laser matrices at high pump fluence values (>1 J/cm2) were also analyzed. These effects originate, most likely, from a two-step photothermal mechanism.