Search

Logo detection is one of the crucial branches in computer vision due to various real-world applications, such as automatic logo detection and recognition, intelligent transportation, and trademark infringement detection. Compared with traditional handcrafted-feature-based methods, deep learning-based convolutional neural networks (CNNs) can learn both low-level and high-level image features. Recent decades have witnessed the great feature representation capabilities of deep CNNs and their variants, which have been very good at discovering intricate structures in high-dimensional data and are thereby applicable to many domains including logo detection. However, logo detection remains challenging, as existing detection methods cannot solve well the problems of a multiscale and large aspect ratios. In this paper, we tackle these challenges by developing a novel long-range dependence involutional network (LDI-Net). Specifically, we designed a strategy that combines a new operator and a self-attention mechanism via rethinking the intrinsic principle of convolution called long-range dependence involution (LD involution) to alleviate the detection difficulties caused by large aspect ratios. We also introduce a multilevel representation neural architecture search (MRNAS) to detect multiscale logo objects by constructing a novel multipath topology. In addition, we implemented an adaptive RoI pooling module (ARM) to improve detection efficiency by addressing the problem of logo deformation. Comprehensive experiments on four benchmark logo datasets demonstrate the effectiveness and efficiency of the proposed approach. [ABSTRACT FROM AUTHOR]

Conventional Staudinger reductions of organic azides are sluggish with aryl or bulky aliphatic azides. In addition, Staudinger reduction usually requires a large excess of water to promote the decomposition of the aza-ylide intermediate into phosphine oxide and amine products. To overcome the challenges above, we designed a novel triaryl phosphine reagent 2c with an ortho-SO2NH2 substituent. Herein, we report that such phosphine reagents are able to mediate the Staudinger reduction of both aryl and alkyl azides in either anhydrous or wet solvents. Good to excellent yields were obtained in all cases (even at a diluted concentration of 0.01 M). The formation of B-TAP, a cyclic aza-ylide, instead of phosphine oxide, eliminates the requirement of water in the Staudinger reduction. In addition, computational studies disclose that the intramolecular protonation of the aza-ylide by the ortho-SO2NH2 group is kinetically favorable and responsible for the acceleration of Staudinger reduction of the aryl azides. [ABSTRACT FROM AUTHOR]

Radar-based classification of human activities and gait have attracted significant attention with a large number of approaches proposed in terms of features and classification algorithms. A common approach in activity classification attempts to find the algorithm (features plus classifier) that can deal with multiple activities analysed in one study such as walking, sitting, drinking and crawling. However, using the same set of features for multiple activities can be suboptimal per activity and not take into account the diversity of kinematic movements that could be captured by diverse features. In this paper, we propose a hierarchical classification approach that uses a large variety of features including but not limited to energy features like entropy and energy curve, physical features like centroid and bandwidth, image-based features like skewness extracted from multiple radar data domains. Feature selection is used at each step of the hierarchical model to select the best set of features to discriminate the target activity from the others, showing improvements with respect to the more conventional approach of using a multiclass model. The proposed approach is validated on a large dataset with 1078 recorded samples of varying length from 5 s to 10 s of experimental data, yielding 95.4% accuracy to classify six activities. The approach is also validated on a personnel recognition task to identify individual subjects from their walking gait, yielding 83.7% accuracy for ten subjects and 68.2% for a significantly larger group of subjects, i.e., 60 people. [ABSTRACT FROM AUTHOR]

Background: Admission to hospital introduces risks for people with Parkinson's disease in maintaining continuity of their highly individualized medication regimens, which increases their risk of medication errors. This is of particular concern as omitted medications and irregular dosing can cause an immediate increase in an individual's symptoms as well as other adverse outcomes such as swallowing difficulties, aspiration pneumonia, frozen gait and even potentially fatal neuroleptic malignant type syndrome. Objective: To determine the occurrence and identify factors that contribute to Parkinson's medication errors in Australian hospitals. Methods: A retrospective discharge diagnosis code search identified all admissions for people with Parkinson's disease to three tertiary metropolitan hospitals in South Australia, Australia over a 3-year period. Of the 405 case notes reviewed 351 admissions met our inclusion criteria. Results: Medication prescribing (30.5%) and administration (85%) errors during admission were extremely common, with the most frequent errors related to administration of levodopa preparations (83%). A higher levodopa equivalent dosage, patients with a modified swallowing status or nil by mouth order during admission, and patients who did not have a pharmacist led medication history within 24 hours of admission had significantly higher rates of medication errors. Conclusions: This study identified 3 major independent factors that increased the risk of errors during medication management for people with Parkinson's disease during hospitalization. Thus, targeting these areas for preventative interventions have the greatest chance of producing a clinically meaningful impact on the number of hospital medication errors occurring in the Parkinson's population. [ABSTRACT FROM AUTHOR]