Search

Few-shot human-object interaction (FS-HOI) recognition aims at inferring new interactions between human actions and surrounding objects merely with a few available instances. It is beneficial to alleviate the long-tail and combinatorial explosion problems in human-object interaction (HOI). Nevertheless, the existing FS-HOI methods only focus on modeling the relationships between labeled samples and unlabeled samples in the Euclidean domain, which neglects the rich relational structures of the visual information among labeled samples and between human actions and objects. Accordingly, we tackle the few-shot HOI task in the non-Euclidean domain and present a graph-based model, namely, task-oriented high-order context graph network (THCG-Net). It contains a task attention module (TA-Module) and a high-order context graph module (HG-Module). In TA-Module, an attention mechanism is designed by utilizing task information to build a task-oriented space, in which the discriminative information for the current task (episode) is captured by embedding the visual features into the task-oriented space. The HG-Module is proposed to construct a task-level graph and takes the context information as high-order knowledge, which provides discriminative guidance for propagating visual information. It captures the discriminability among different categories while highlights the commonality of related categories adaptively, which effectively transfers knowledge to related categories. Extensive experimental results on two benchmark datasets, HICO-FS and TUHOI-FS, are provided. It demonstrates that our THCG-Net significantly outperforms the state-of-the-art approaches, which proves its impressive effectiveness in recognizing various human actions and surrounding objects in few-shot scenarios. [ABSTRACT FROM AUTHOR]

Few-shot learning (FSL) for human–object interaction (HOI) aims at recognizing various relationships between human actions and surrounding objects only from a few samples. It is a challenging vision task, in which the diversity and interactivity of human actions result in great difficulty to learn an adaptive classifier to catch ambiguous interclass information. Therefore, traditional FSL methods usually perform unsatisfactorily in complex HOI scenes. To this end, we propose dynamic graph-in-graph networks (DGIG-Net), a novel graph prototypes framework to learn a dynamic metric space by embedding a visual subgraph to a task-oriented cross-modal graph for few-shot HOI. Specifically, we first build a knowledge reconstruction graph to learn latent representations for HOI categories by reconstructing the relationship among visual features, which generates visual representations under the category distribution of every task. Then, a dynamic relation graph integrates both reconstructible visual nodes and dynamic task-oriented semantic information to explore a graph metric space for HOI class prototypes, which applies the discriminative information from the similarities among actions or objects. We validate DGIG-Net on multiple benchmark datasets, on which it largely outperforms existing FSL approaches and achieves state-of-the-art results. [ABSTRACT FROM AUTHOR]

Semantic information provides intra-class consistency and inter-class discriminability beyond visual concepts, which has been employed in Few-Shot Learning (FSL) to achieve further gains. However, semantic information is only available for labeled samples but absent for unlabeled samples, in which the embeddings are rectified unilaterally by guiding the few labeled samples with semantics. Therefore, it is inevitable to bring a cross-modal bias between semantic-guided samples and nonsemantic-guided samples, which results in an information asymmetry problem. To address this problem, we propose a Modal-Alternating Propagation Network (MAP-Net) to supplement the absent semantic information of unlabeled samples, which builds information symmetry among all samples in both visual and semantic modalities. Specifically, the MAP-Net transfers the neighbor information by the graph propagation to generate the pseudo-semantics for unlabeled samples guided by the completed visual relationships and rectify the feature embeddings. In addition, due to the large discrepancy between visual and semantic modalities, we design a Relation Guidance (RG) strategy to guide the visual relation vectors via semantics so that the propagated information is more beneficial. Extensive experimental results on three semantic-labeled datasets, i.e., Caltech-UCSD-Birds 200-2011, SUN Attribute Database and Oxford 102 Flower, have demonstrated that our proposed method achieves promising performance and outperforms the state-of-the-art approaches, which indicates the necessity of information symmetry. [ABSTRACT FROM AUTHOR]

Despite the success of watermarking technique for protecting depth image-based rendering (DIBR) 3-D videos, existing methods still can hardly ensure the robustness against geometric attacks, lossless video quality, and distinguishability between different videos simultaneously. In this article, we propose a novel zero-watermarking scheme to address this challenge. Specifically, we design CT-SVD features to ensure both distinguishability and robustness against signal processing and DIBR conversion attacks. In addition, a logistic–logistic chaotic system is utilized to encrypt features for the enhanced security. Moreover, a rectification mechanism based on salient map detection and SIFT matching is designed to resist geometric attacks. Finally, we establish an attention-based fusion mechanism to explore the complementary robustness of rectified and unrectified features. Experimental results demonstrate that our proposed method outperforms the existing schemes in terms of losslessness, distinguishability, and robustness against geometric attacks. [ABSTRACT FROM AUTHOR]

A pattern- and polarization-reconfigurable dielectric resonator antenna (DRA) is presented. It deploys a conventional cylindrical dielectric resonator (DR) with a new switchable feeding network. The feeding network has four slots that are etched in the ground. By tuning the phase distributions in these slots, three DR modes (the TM01δ and two degenerated HEM113 modes) are excited. The former has an omnidirectional radiation pattern, whereas the latter two have boresight patterns with orthogonal polarizations (x- and y-polarization). By tuning the height-to-radius ratio, these modes can be designed to operate at the same frequency, thus enabling both the pattern and polarization reconfigurability. A prototype of the reconfigurable DRA working in 2.4 GHz is fabricated and measured to verify ideas. Reasonable agreement between measurements and simulations is obtained. Our design provides both pattern and polarization reconfigurability, with good efficiency in the band of interest. [ABSTRACT FROM AUTHOR]

A compact horizontally polarized omnidirectional cylindrical dielectric resonator antenna (DRA) excited in the TE $_{011+\delta }$ mode is presented. It deploys a planar feed which comprises a cross-shaped feed line, four coupled strips, and four curved arms with end-shorted stubs. The TE $_{011+\delta }$ mode of the DRA is excited by the curved arms. Its intrinsically narrow impedance bandwidth is enhanced by using the coupled strips, and the cross-polar field is suppressed by the end-shorted stubs. The antenna has a measured −10 dB impedance bandwidth of 18.1%, which is much wider than the conventional bandwidths of 4.5%–7.4%. Its cross-polar level is lower than the co-polar level by 15 dB. This basic DRA is used to design the first polarization reconfigurable omnidirectional DRA with diodes. By controlling the states of the diodes, the DRA can be switched between TE $_{011+\delta }$ and TM $_{01\delta }$ modes to obtain horizontal and vertical polarizations, respectively. To verify the idea, a prototype was fabricated and tested. Reasonable agreement between the measured and simulated results is obtained. The measured impedance bandwidths of the TE $_{011+\delta }$ and TM $_{01\delta }$ modes are 16.8% (2.18–2.58 GHz) and 16.0% (2.25–2.64 GHz), respectively. The antenna has stable omnidirectional radiation patterns in both states [ABSTRACT FROM AUTHOR]

Extreme instance imbalance among categories and combinatorial explosion make the recognition of Human-Object Interaction (HOI) a challenging task. Few studies have addressed both challenges directly. Motivated by the success of few-shot learning that learns a robust model from a few instances, we formulate HOI as a few-shot task in a meta-learning framework to alleviate the above challenges. Due to the fact that the intrinsical characteristic of HOI is diverse and interactive, we propose a Semantic-guided Attentive Prototypes Network (SAPNet) framework to learn a semantic-guided metric space where HOI recognition can be performed by computing distances to attentive prototypes of each class. Specifically, the model generates attentive prototypes guided by the category names of actions and objects, which highlight the commonalities of images from the same class in HOI. In addition, we design two alternative prototypes calculation methods, i.e., Prototypes Shift (PS) approach and Hallucinatory Graph Prototypes (HGP) approach, which explore to learn a suitable category prototypes representations in HOI. Finally, in order to realize the task of few-shot HOI, we reorganize 2 HOI benchmark datasets with 2 split strategies, i.e., HICO-NN, TUHOI-NN, HICO-NF, and TUHOI-NF. Extensive experimental results on these datasets have demonstrated the effectiveness of our proposed SAPNet approach. [ABSTRACT FROM AUTHOR]

A filtering dielectric resonator (DR) antenna (DRA) that does not require any filtering circuits is investigated. It deploys a rectangular DR and three slots fabricated in the ground plane. The antenna is excited in the fundamental TEy1Δ1 DRA mode and two slot modes. By making use of these three modes, two tunable radiation nulls near the edges of the passband can be obtained, giving filtering effects without the need for any filtering circuits. In this article, the design, with minor modifications of the slots, is extended to a frequency-reconfigurable filtering DRA by using varactors. Its stopband can be extended to the third harmonic. By varying the voltages of the varactors, the antenna has continuous frequency agility of 27.7% (2.15–2.84 GHz), with almost the same harmonics suppression level. The antenna has a gain of ~5.0 dBi and a rejection level of over 16.1 dB across the stopband. Its total antenna efficiency in the passband is ~81%, whereas it is generally lower than the average value of 1.8% in the stopband (up to 7.0 GHz). Compared with the reported frequency-reconfigurable filtering antennas, our design has better frequency agility, more stable filtering response, and weaker radiation in the stopband. [ABSTRACT FROM AUTHOR]

A dual-frequency continuous-wave (DF-CW) radar is a cost-effective device for multiple target tracking in indoor environments. The traditional DF-CW radar adopts a quasi-monostatic structure to acquire the target range and direction of arrival (DOA), to position the target. However, the traditional DF-CW radar usually suffers from the drawbacks of low positioning accuracy, biased target information, and severe ambiguous frequency interference. For overcoming these drawbacks, the traditional DF-CW radar structure is modified, and an improved target tracking algorithm, which is based only on the range measurements, is proposed in this article. Compared with the traditional DF-CW radar-based detection, the proposed algorithm can achieve not only higher positioning accuracy, reveal more abundant and precise target information, but can also help to suppress the ambiguous frequency interference, which is of great importance in the multiple target tracking scenarios. Besides, the proposed DF-CW radar can also effectively reduce the coupling interference, thus contributing to the improvement of the dynamic range of the radar. A series of experiments are conducted to demonstrate the effectiveness of the proposed algorithm, and the obtained results illustrate a better performance of the proposed algorithm compared with the traditional algorithms. [ABSTRACT FROM AUTHOR]

Tethered space net robot (TSNR) has been proposed as a promising method for active space debris capturing and removal since space debris have posed a big threat to operational satellites and astronauts. After contacting the debris, the net cannot naturally wrap around the debris, especially when the debris does not contact the net in the center. Thus, it is worth investigating the net closing control of TSNR. Two requirements should be satisfied in the net closing control algorithm. First, the control algorithm should make the net close around the debris in a very short time to avoid the escape of the debris. Second, the control input should be continuous to avoid the oscillations of net tether caused by high-frequency control instructions. In this article, based on the two requirements, an impulsive adaptive super-twisting sliding mode control scheme is proposed to complete the net closing action. The control scheme can drive the state vectors convergence to the origin in a short time and keep it there thereafter. The numerical simulation results illustrate the efficiency of the proposed control algorithm, in which the two requirements are all satisfied. [ABSTRACT FROM AUTHOR]

Glaucoma is a chronic eye disease that leads to irreversible vision loss. The Cup-to-Disc Ratio (CDR) serves as the most important indicator for glaucoma screening and plays a significant role in clinical screening and early diagnosis of glaucoma. In general, obtaining CDR is subjected to measuring on manually or automatically segmented optic disc and cup. Despite great efforts have been devoted, obtaining CDR values automatically with high accuracy and robustness is still a great challenge due to the heavy overlap between optic cup and neuroretinal rim regions. In this paper, a direct CDR estimation method is proposed based on the well-designed semi-supervised learning scheme, in which CDR estimation is formulated as a general regression problem while optic disc/cup segmentation is cancelled. The method directly regresses CDR value based on the feature representation of optic nerve head via deep learning technique while bypassing intermediate segmentation. The scheme is a two-stage cascaded approach comprised of two phases: unsupervised feature representation of fundus image with a convolutional neural networks (MFPPNet) and CDR value regression by random forest regressor. The proposed scheme is validated on the challenging glaucoma dataset Direct-CSU and public ORIGA, and the experimental results demonstrate that our method can achieve a lower average CDR error of 0.0563 and a higher correlation of around 0.726 with measurement before manual segmentation of optic disc/cup by human experts. Our estimated CDR values are also tested for glaucoma screening, which achieves the areas under curve of 0.905 on dataset of 421 fundus images. The experiments show that the proposed method is capable of state-of-the-art CDR estimation and satisfactory glaucoma screening with calculated CDR value. [ABSTRACT FROM AUTHOR]

A gain-enhanced tri-band microstrip antenna using higher order modes was analyzed and designed in this communication. It first proved in theory that by manipulating the equivalent magnetic currents’ (EMCs) relative magnitudes, the higher order modes’ radiation patterns could be converted from conical to broadside. Their zenithal directivities would therefore greatly increase. The relationships between the EMCs’ relative magnitudes and the directivities were studied. Then, those analyses were applied to design a gain-enhanced tri-band microstrip antenna using three higher order modes. A simpler one-port differential feeding and several shorting pins and slots were used to control the antenna’s EMCs to improve its radiation patterns. They could further increase the antenna’s gains and suppress its sidelobe levels and cross-polarizations. The proposed antenna worked at 3.9, 4.95, and 5.9 GHz with consistent broadside radiation patterns and the same polarization. Their measured gains were about 10.0, 10.4, and 10.2 dBi, respectively. [ABSTRACT FROM AUTHOR]

A dual-band sector ring microstrip antenna operating at the TM11 and TM02 modes is analyzed and then designed in this communication. Unlike previous analyses that use only cavity model or transmission line, a simpler analysis using both models is proposed. Accordingly, several closed equations are derived. Such equations determine the relationship among three important units of construction, namely the radii, the angle, and its operating frequencies, and with the operating factors in a dual-band sector ring patch. The errors between the proposed method and the cavity model are acceptable. The field and radiation patterns of the sector ring patch are also discussed. In addition, these analyses and equations are applied to design a dual-band sector ring microstrip antenna for WLAN. A novel coupling sector ring patch is then proposed to improve the performance of the antenna. The experimental results are in good agreement with the analyses and the simulations. The impedance bandwidths of the antenna are about 40 and 900 MHz with consistent broadside radiation patterns at 2.4 and 5.5 GHz band. Additionally, their measured average gains are about 5.1 and 6.3 dBi, respectively. [ABSTRACT FROM PUBLISHER]

In this letter, a localization algorithm, which combines Hough transform and Doppler processing, is proposed for Doppler radar human sensing applications. The Hough transform is first applied to extract the interested target components and real-time estimate their instantaneous frequencies (IFs). Then, based on the IF estimation result, the target movement trajectories are synthesized by Doppler processing. To improve the detection accuracy and robustness, a generalized linear model is proposed for Hough transform, which can achieve highly dimensional frequency fitting to compensate the nonlinear error, meanwhile maintaining a low level of computational complexity for real-time processing. Experimental results are presented to illustrate the performance of the proposed algorithm. [ABSTRACT FROM PUBLISHER]