Three-Dimensional Reconstruction of Cerebrovascular and Algorithm Realization

Objective. In the three-dimensional reconstruction of CT cerebrovascular medical image registration, a new optimization algorithm based on the relative position information between the contours of various blood vessels in the image is proposed. Methods. Using the rule that the center of gravity of the vascular tissue structure on the series of slices has continuity, find the registration relationship between the contours of the vessels in the two adjacent slices. Because the shape of cerebrovascular contour is relatively symmetrical, its center of gravity is slightly away from its geometric center. Therefore, the geometric center is used to replace the center of gravity, and the “mass” of each contour is calculated according to the area of each contour to achieve the registration of the blood vessel contour. Results. The method has the characteristics of global optimization and stronger robustness. Conclusion. The cerebrovascular image obtained by this method is more realistic and can be used for the import of various software, simulation training, and later research, which provides an effective method for preoperative simulation of cerebrovascular intervention surgery.

Automatic Roadblock Identification Algorithm for Unmanned Vehicles Based on Binocular Vision

In order to improve the accuracy of automatic obstacle recognition algorithm for driverless vehicles, an automatic obstacle recognition algorithm for driverless vehicles based on binocular vision is constructed. Firstly, the relevant parameters of the camera are calibrated around the new car coordinate system to determine the corresponding obstacle position of the vehicle. At the same time, the three-dimensional coordinates of obstacle points are obtained by binocular matching method. Then, the left and right cameras are used to capture the feature points of obstacles in the image to realize the recognition of obstacles. Finally, the experimental results show that for obstacle 1, the recognition error of the algorithm is 0.03 m; for obstacle 2, the recognition error is 0.02 m; for obstacle 3, the recognition error is 0.01 m. The algorithm has small recognition error. The vehicle coordinate system is added in the camera calibration process, which can accurately measure the relative position information between the vehicle and the obstacle.

Distributed Event-Triggered Approach for Multi-Agent Formation Based on Cooperative Localization with Mixed Measurements

This paper concentrates on multi-agent formation control problems under mixed measurements of distance and bearing. Towards this objective, a distributed event-triggered estimation-based control framework is developed such that only at necessary time instants, the event for estimation (namely, cooperative localization among a subgroup of agents) is triggered to recover relative position information by utilizing a mixed distance and bearing measurements from different agents. Firstly, it is shown by using the stiffness theory that a subgroup of agents are capable of recovering relative position information if a sufficient number of independent distance and range measurements are available. Secondly, a distributed event-triggered mechanism is presented for achieving an affine formation control, which can be implemented in an asynchronous manner and also ensures Zeno-free behavior. Simulation studies are provided to demonstrate the effective performance of the proposed approach.

RoRePo: Detecting the role information and relative position information for contexts in multi-turn dialogue generation

In multi-turn dialogue generation, dialogue contexts have been shown to have an important influence on the reasoning of the next round of dialogue. A multi-turn dialogue between two people should be able to give a reasonable response according to the relevant context. However, the widely used hierarchical recurrent encoder-decoder model and the latest model that detecting the relevant contexts with self-attention are facing the same problem. Their given response doesn’t match the identity of the current speaker, which we call it role ambiguity. In this paper, we propose a new model, named RoRePo, to tackle this problem by detecting the role information and relative position information. Firstly, as a part of the decoder input, we add a role embedding to identity different speakers. Secondly, we incorporate self-attention mechanism with relative position representation to dialogue context understanding. Besides, the design of our model architecture considers the influence of latent variables in generating more diverse responses. Experimental results of our evaluations on the DailyDialog and DSTC7_AVSD datasets show that our proposed model advances in multi-turn dialogue generation.

Aspect-Level Sentiment Analysis Based on Position Features Using Multilevel Interactive Bidirectional GRU and Attention Mechanism

The aim of aspect-level sentiment analysis is to identify the sentiment polarity of a given target term in sentences. Existing neural network models provide a useful account of how to judge the polarity. However, context relative position information for the target terms is adversely ignored under the limitation of training datasets. Considering position features between words into the models can improve the accuracy of sentiment classification. Hence, this study proposes an improved classification model by combining multilevel interactive bidirectional Gated Recurrent Unit (GRU), attention mechanisms, and position features (MI-biGRU). Firstly, the position features of words in a sentence are initialized to enrich word embedding. Secondly, the approach extracts the features of target terms and context by using a well-constructed multilevel interactive bidirectional neural network. Thirdly, an attention mechanism is introduced so that the model can pay greater attention to those words that are important for sentiment analysis. Finally, four classic sentiment classification datasets are used to deal with aspect-level tasks. Experimental results indicate that there is a correlation between the multilevel interactive attention network and the position features. MI-biGRU can obviously improve the performance of classification.

Self-Attention Enhanced Selective Gate with Entity-Aware Embedding for Distantly Supervised Relation Extraction

Distantly supervised relation extraction intrinsically suffers from noisy labels due to the strong assumption of distant supervision. Most prior works adopt a selective attention mechanism over sentences in a bag to denoise from wrongly labeled data, which however could be incompetent when there is only one sentence in a bag. In this paper, we propose a brand-new light-weight neural framework to address the distantly supervised relation extraction problem and alleviate the defects in previous selective attention framework. Specifically, in the proposed framework, 1) we use an entity-aware word embedding method to integrate both relative position information and head/tail entity embeddings, aiming to highlight the essence of entities for this task; 2) we develop a self-attention mechanism to capture the rich contextual dependencies as a complement for local dependencies captured by piecewise CNN; and 3) instead of using selective attention, we design a pooling-equipped gate, which is based on rich contextual representations, as an aggregator to generate bag-level representation for final relation classification. Compared to selective attention, one major advantage of the proposed gating mechanism is that, it performs stably and promisingly even if only one sentence appears in a bag and thus keeps the consistency across all training examples. The experiments on NYT dataset demonstrate that our approach achieves a new state-of-the-art performance in terms of both AUC and top-n precision metrics.

Perceptual Information Supports Transfer of Learning in Coordinated Rhythmic Movement

In this paper, we trained people to produce 90° mean relative phase using task-appropriate feedback and investigated whether and how that learning transfers to other coordinations. Past work has failed to find transfer of learning to other relative phases, only to symmetry partners (identical coordinations with reversed lead-lag relationships) and to other effector combinations. However, that research has all trained people using transformed visual feedback (visual metronomes, Lissajous feedback) which removes the relative motion information typically used to produce various coordinations (relative direction, relative position; Wilson & Bingham, 2008) . Coordination feedback (Wilson, Snapp-Childs, Coats, & Bingham, 2010) preserves that information and we have recently shown that relative position supports transfer of learning between unimanual and bimanual performance of 90° (Snapp-Childs, Wilson, & Bingham, 2015). Here we ask whether that information can support the production of other relative phases. We found large, asymmetric transfer of learning bimanual 90° to bimanual 60° and 120°, supported by perceptual learning of relative position information at 90°. For learning to transfer, the two tasks must overlap in some critical way; this is additional evidence that this overlap must be informational. We discuss the results in the context of an ecological, task dynamical approach to understanding the nature of perception-action tasks.