Due to its storage and retrieval efficiency, cross-modal hashing (CMH) has been widely used for cross-modal similarity search in many multimedia applications. According to the training strategy, ...existing CMH methods can be mainly divided into two categories: relaxation-based continuous methods and discrete methods. In general, the training of relaxation-based continuous methods is faster than that of discrete methods, but the accuracy of relaxation-based continuous methods is not satisfactory. On the contrary, the accuracy of discrete methods is typically better than that of the relaxation-based continuous methods, but the training of discrete methods is very time-consuming. In this paper, we propose a novel CMH method, called Discrete Latent Factor model-based cross-modal Hashing (DLFH), for cross modal similarity search. DLFH is a discrete method which can directly learn the binary hash codes for CMH. At the same time, the training of DLFH is efficient. Experiments show that the DLFH can achieve significantly better accuracy than existing methods, and the training time of DLFH is comparable to that of the relaxation-based continuous methods which are much faster than the existing discrete methods.
Hashing has been widely used for large-scale search due to its low storage cost and fast query speed. By using supervised information, supervised hashing can significantly outperform unsupervised ...hashing. Recently, discrete supervised hashing and feature learning based deep hashing are two representative progresses in supervised hashing. On one hand, hashing is essentially a discrete optimization problem. Hence, utilizing supervised information to directly guide discrete (binary) coding procedure can avoid sub-optimal solution and improve the accuracy. On the other hand, feature learning based deep hashing, which integrates deep feature learning and hash-code learning into an end-to-end architecture, can enhance the feedback between feature learning and hash-code learning. The key in discrete supervised hashing is to adopt supervised information to directly guide the discrete coding procedure in hashing. The key in deep hashing is to adopt the supervised information to directly guide the deep feature learning procedure. However, most deep supervised hashing methods cannot use the supervised information to directly guide both discrete (binary) coding procedure and deep feature learning procedure in the same framework. In this paper, we propose a novel deep hashing method, called deep discrete supervised hashing (DDSH). DDSH is the first deep hashing method which can utilize pairwise supervised information to directly guide both discrete coding procedure and deep feature learning procedure and thus enhance the feedback between these two important procedures. Experiments on four real datasets show that DDSH can outperform other state-of-the-art baselines, including both discrete hashing and deep hashing baselines, for image retrieval.
Deep Cross-Modal Hashing Qing-Yuan Jiang; Wu-Jun Li
2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR),
2017-July
Conference Proceeding
Due to its low storage cost and fast query speed, cross-modal hashing (CMH) has been widely used for similarity search in multimedia retrieval applications. However, most existing CMH methods are ...based on hand-crafted features which might not be optimally compatible with the hash-code learning procedure. As a result, existing CMH methods with hand-crafted features may not achieve satisfactory performance. In this paper, we propose a novel CMH method, called deep cross-modal hashing (DCMH), by integrating feature learning and hash-code learning intothe same framework. DCMH is an end-to-end learning framework with deep neural networks, one for each modality, to perform feature learning from scratch. Experiments on three real datasets with image-text modalities show that DCMH can outperform other baselines to achieve the state-of-the-art performance in cross-modal retrieval applications.
Through the synergies of a heterogeneous synthetic approach and a homogeneous synthetic methodology, N‐containing organic compounds can be synthesized via activated N‐containing species prepared from ...N2 gas and suitable carbon sources. From N2, carbon, and LiH, we have previously succeeded in the high‐yield preparation of Li2CN2 as the activated N‐containing species. In this work, we applied Li2CN2 as a novel synthetic synthon for constructing N‐containing organic compounds. A series of reaction models, including a substitution reaction, cycloaddition reaction, and transition metal‐catalyzed coupling reaction, were successfully performed using Li2CN2 under mild conditions. Various valuable cyanamides, carbodiimides, N‐aryl cyanamides and 1,2,4‐triazole derivatives were readily synthesized in moderate to excellent yields. With this method, the 15N‐labeled products, including oxazolidine derivatives with anti‐cancer activity, could also be facilely prepared from 15N2 gas.
A strategy for the conversion of inert N2 into various nitrogen‐containing organic compounds is presented. The activated N‐containing species Li2CN2 is successfully prepared in high yields from N2 gas, C and LiH. Li2CN2 can be transformed into a range of N‐containing organic compounds, including cyanamides, carbodiimides, N‐aryl cyanamides and 1,2,4‐triazole derivatives. 15N‐labeled products, including oxazolidine derivatives, have also been prepared from 15N2 gas.
Due to its successful application in recommender systems, collaborative filtering (CF) has become a hot research topic in data mining and information retrieval. In traditional CF methods, only the ...feedback matrix, which contains either explicit feedback (also called ratings) or implicit feedback on the items given by users, is used for training and prediction. Typically, the feedback matrix is sparse, which means that most users interact with few items. Due to this sparsity problem, traditional CF with only feedback information will suffer from unsatisfactory performance. Recently, many researchers have proposed to utilize auxiliary information, such as item content (attributes), to alleviate the data sparsity problem in CF. Collaborative topic regression (CTR) is one of these methods which has achieved promising performance by successfully integrating both feedback information and item content information. In many real applications, besides the feedback and item content information, there may exist relations (also known as networks) among the items which can be helpful for recommendation. In this paper, we develop a novel hierarchical Bayesian model called Relational Collaborative Topic Regression (RCTR), which extends CTR by seamlessly integrating the user-item feedback information, item content information, and network structure among items into the same model. Experiments on real-world datasets show that our model can achieve better prediction accuracy than the state-of-the-art methods with lower empirical training time. Moreover, RCTR can learn good interpretable latent structures which are useful for recommendation.
A new palladium‐catalyzed reductive 5+1 cycloaddition of 3‐acetoxy‐1,4‐enynes with CO, enabled by hydrosilanes, has been developed for delivering valuable functionalized phenols. This methodology ...employs hydrosilanes as the external reagent to facilitate the 5+1 carbonylative benzannulation. The reaction is a conceptually and mechanistically novel carbonylative cycloaddition route for the construction of substituted phenols, through the formation of four new chemical bonds, with excellent functional‐group tolerance.
Pd & CO: Employing reductive palladium catalysis enables a new 5+1 carbonylative benzannulation of 3‐acetoxy‐1,4‐enynes with CO and hydrosilanes. This reaction is facilitated by hydrosilanes, and allows straightforward and practical access to functionalized phenols with excellent functional‐group tolerance and high selectivity.
We report the isolation and characterization of a novel bat coronavirus which is much closer to the severe acute respiratory syndrome coronavirus (SARS-CoV) in genomic sequence than others previously ...reported, particularly in its S gene. Cell entry and susceptibility studies indicated that this virus can use ACE2 as a receptor and infect animal and human cell lines. Our results provide further evidence of the bat origin of the SARS-CoV and highlight the likelihood of future bat coronavirus emergence in humans.
A novel multicomponent sulfonylation of alkenes is described for the assembly of various β-substituted arylsulfones using cheap and easily available K2S2O5 as a sulfur dioxide source. Of note, the ...procedure does not need any extra oxidants and metal catalysts and exhibits a relatively wide substrate scope and good functional group compatibility. Mechanistically, an initial arylsulfonyl radical is formed involving the insertion of sulfur dioxide with aryl diazonium salt, followed by alkoxyarylsulfonylation or hydroxysulfonylation of alkenes.
Terrible triad injury of the elbow (TTIE), comprising elbow dislocation with radial head and coronoid process fracture, is notoriously challenging to treat and has typically been associated with ...complications and poor outcomes. The objective of this systematic review was to summarize the most recent available evidence regarding functional outcomes and complications following surgical management of TTIE.
Medline, EMBASE, Cochrane Library, and Google Scholar were searched to identify relevant studies, which were included if they were retrospective or prospective in design, involved participants who had TTIE, and were published in English. Outcomes of interest were functional outcomes and complications.
Sixteen studies, involving 312 patients, were included in the systematic review. Mean follow up after surgery was typically 25 to 30 months. Mean Mayo elbow performance scores ranged from 78 to 95. Mean Broberg-Morrey scores ranged from 76 to 90. Mean DASH scores ranged from 9 to 31. The proportion of patients who required reoperation due to complications ranged from 0 to 54.5% (overall = 70/312 22.4%). Most of these complications were related to hardware fixation problems, joint stiffness, joint instability, and ulnar neuropathy. The most common complications that did not require reoperation were heterotopic ossification (39/312 12.5% patients) and arthrosis (35/312 11.2% patients).
The results of this systematic review indicate that functional outcomes after surgery for TTIE are generally satisfactory and that complications are common. Further research is warranted to determine which surgical techniques optimize functional outcomes and reduce the risk of complications.
•A 2D vision-based non-contact displacement sensor system is constructed.•The space coordinates of targets are acquired by DLT and registration approaches.•The system performances are quite stable in ...different camera orientations.•The system flexibility is improved.
In the class of not-contact sensors, the techniques of vision-based displacement estimation enable one to gather dense global measurements of static deformation as well as of dynamic response. They are becoming more and more available thanks to the ongoing technology developments. In this work, a vision system, which takes advantage of fast-developing digital image processing and computer vision technologies and provides high sample rate, is implemented to monitor the 2D plane vibrations of a reduced scale frame mounted on a shaking table as available in a laboratory. The physical meanings of the camera parameters, the trade-off between the system resolution and the field-of-view, and the upper limitation of marker density are discussed. The scale factor approach, which is widely used to convert the image coordinates measured by a vision system in the unit of pixels into space coordinates, causes a poor repeatability of the experiment, an unstable experiment precision, and therefore a global poor flexibility. To overcome these problems, two calibrations approaches are introduced: registration and direct linear transformation. Based on the constructed vision-based displacement measurement system, several experiments are carried out to monitor the motion of a scale-reduced model on which dense markers are glued. The experiment results show that the proposed system can capture and successfully measure the motion of the laboratory model within the required frequency band.