Search

Introducción. La implantología oral representa una opción importante en el tratamiento de los pacientes con pérdida dental total y parcial. El estudio a largo plazo muestra la evaluación de los pacientes tratados mediante carga precoz de implantes con conexión interna y superficie arenada y grabada. Métodos. 40 pacientes con pérdidas dentales fueron tratados con implantes Frontier GMI ® con conexión interna y superficie arenada y grabada. Los implantes fueron cargados tras un periodo de tiempo de 6 semanas en la mandíbula y 8 semanas en el maxilar superior. Los hallazgos clínicos (implantológicos y prostodóncicos) se han seguido durante al menos 6 años. Resultados. 103 implantes fueron insertados en ambos maxilares, 46 implantes (44,7%) en el maxilar superior y 57 implantes (55,3%) en la mandíbula, para su rehabilitación prostodóncica. 37 implantes (35,9%) fueron insertados en el sector anterior y 66 implantes (64,1%) en el sector posterior. 70 implantes (68%) de forma sumergida (2 cirugías) y 33 implantes (32%) de forma no sumergida (una cirugía). Des pués de un seguimiento clínico de 92,2 meses, los resultados indican una supervivencia de los implantes del 96,2%; ya que se perdieron 4 implantes. La pérdida ósea marginal fué de 0,99 ± 0,84 mm. Las restau raciones prostodóncicas incluyeron 31 coronas unitarias, 15 puentes fijos, 5 sobredentaduras con bolas, 2 rehabilitaciones totales fijas y una rehabilitación completa híbrida. Las complicaciones tardías fueron 11 implantes (10,7%) con periimplantitis y 6 prótesis (11,1%) con complicaciones protésicas. Conclusiones. Los hallazgos clínicos del estudio indican que el tratamiento con carga precoz mediante prótesis de los implantes con conexión interna y superficie arenada y grabada, representa una terapéutica odontológica con éxito. Introduction. Implant dentistry constitute an important option in the prosthodontic treatment of patients with partial and total tooth loss. This long-term study reports the evaluation of patients treated by early loading of implants with internal connection and sandblasted-acidetched surface. Methods. 40 patients with tooth loss were treated with Frontier GMI ® sandblasted and acid-etched surfa ce implants. Implants were loaded after a healing free-loading period of 6 weeks (mandible) and 8 weeks (maxilla). Clinical findings (implants and prosthodontics) were followed during at least 6 years. Results. 103 implants were inserted (46 maxillary (44.7%), and 57 mandibular (55.3%)) for prosthodontic rehabilitation. 37 implants (35.9%) were inserted in anterior sites and 66 implant (64.1%) in posterior sites. Seventy implants (68%) were placed submerged (two stages) while that 33 implants (32%) were placed nonsubmerged (one stage). After at a mean follow-up of 92.2 months, clinical results indicate a survival rate of implants of 96.2%. Four implants were lost during the treatment. Mean marginal bone loss were 0.99 ± 0.84 mm. Prosthodontic restorations included 31 single crowns, 15 fixed bridges, 5 overdentures, 2 fixed totally rehabilitation and an hybrid full rehabilitation. Delayed complications include 11 implants (10.7%) with peri-implantitis and 6 prostheses (11.1%) with technical complications. Conclusions. Clinical results of this study indicate that prosthodontic rehabilitation by early loading of internal connection and sandblasted and etched- implants can be a successful dental treatment.

The labeling process within a supervised learning task is usually carried out by an expert, which provides the ground truth (gold standard) for each sample. However, in many real-world applications, we typically have access to annotations provided by crowds holding different and unknown expertise levels. Learning from crowds (LFC) intends to configure machine learning paradigms in the presence of multilabelers, residing on two key assumptions: the labeler's performance does not depend on the input space, and independence among the annotators is imposed. Here, we propose the correlated chained Gaussian processes from the multiple annotators (CCGPMA) approach, which models each annotator's performance as a function of the input space and exploits the correlations among experts. Experimental results associated with classification and regression tasks show that our CCGPMA performs better modeling of the labelers' behavior, indicating that it consistently outperforms other state-of-the-art LFC approaches.

Training a supervised learning model requires a labeled dataset, where the labeling process is usually carried out by an expert to provide the ground truth/gold standard for each sample. However, in many applications, such a gold standard is not available. Instead, several real-world scenarios give us access to annotations provided by crowds, holding different and unknown levels of expertise. Thus, Learning from crowds is a subject undergoing intense study, and its main aim is to manage various machine learning paradigms in the presence of multiple annotators. Most of the state-of-the-art approaches reside on two key assumptions: i) the labeler’s performance does not depend on the input feature space, and ii) independence among the annotators is imposed. Here, we introduce a localized kernel alignment-based annotator relevance analysis (LKAAR) to code each labeler’s expertise as the matching between the feature and label spaces. Namely, LKAAR takes into account inter-annotators dependencies and models labelers’ knowledge as a function of the input feature space. Experimental results devoted to classification, show that LKAAR achieves suitable performances from inconsistent labelers, even if the gold standard is not available.

In a typical supervised learning scenario, it is supposed that there is an oracle who gives the correct label (also known as gold standard or ground truth) for each instance available in the training set. Nevertheless, for many real-world problems, instead of the gold standard, we have access to some annotations (possibly noisy) provided by multiple annotators with different unknown levels of expertise. Then, it is not appropriate to use trivial methods, i.e., majority voting, to estimate the actual label from the annotations due to this way assumes homogeneity in the performance of the labelers. Here, we introduce a new kernel alignment-based annotator relevance analysis–(KAAR) approach to code each annotator expertise as an averaged matching between the input features and the expert labels. So, a new sample label is predicted as a convex combination of classifiers adopting the achieved KAAR-based coding. Experimental results show that our methodology can estimate the performance of annotators even if the gold standard is not available, defeating state-of-the-art techniques.

The increasing popularity of crowdsourcing platforms, i.e., Amazon Mechanical Turk, changes how datasets for supervised learning are built. In these cases, instead of having datasets labeled by one source (which is supposed to be an expert who provided the absolute gold standard), databases holding multiple annotators are provided. However, most state-of-the-art methods devoted to learning from multiple experts assume that the labeler’s behavior is homogeneous across the input feature space. Besides, independence constraints are imposed on annotators’ outputs. This paper presents a regularized chained deep neural network to deal with classification tasks from multiple annotators. The introduced method, termed RCDNN, jointly predicts the ground truth label and the annotators’ performance from input space samples. In turn, RCDNN codes interdependencies among the experts by analyzing the layers’ weights and includes l1, l2, and Monte-Carlo Dropout-based regularizers to deal with the over-fitting issue in deep learning models. Obtained results (using both simulated and real-world annotators) demonstrate that RCDNN can deal with multi-labelers scenarios for classification tasks, defeating state-of-the-art techniques.

Peripheral Nerve Blocking (PNB) is a commonly used technique for performing regional anesthesia and managing pain. PNB comprises the administration of anesthetics in the proximity of a nerve. In this sense, the success of PNB procedures depends on an accurate location of the target nerve. Recently, ultrasound images (UI) have been widely used to locate nerve structures for PNB, since they enable a non-invasive visualization of the target nerve and the anatomical structures around it. However, UI are affected by speckle noise, which makes it difficult to accurately locate a given nerve. Thus, it is necessary to perform a filtering step to attenuate the speckle noise without eliminating relevant anatomical details that are required for high-level tasks, such as segmentation of nerve structures. In this paper, we propose an UI improvement strategy with the use of a pre-image-based filter. In particular, we map the input images by a nonlinear function (kernel). Specifically, we employ a correntropy-based mapping as kernel functional to code higher-order statistics of the input data under both nonlinear and non-Gaussian conditions. We validate our approach against an UI dataset focused on nerve segmentation for PNB. Likewise, our Correntropy-based Pre-Image Filtering (CPIF) is applied as a pre-processing stage to segment nerve structures in a UI. The segmentation performance is measured in terms of the Dice coefficient. According to the results, we observe that CPIF finds a suitable approximation for UI by highlighting discriminative nerve patterns.