Stereo matching algorithms usually consist of four steps, including matching cost calculation, matching cost aggregation, disparity calculation, and disparity refinement. Existing CNN-based methods only adopt CNN to solve parts of the four steps, or use different networks to deal with different steps, making them difficult to obtain the overall optimal solution. In this paper, we propose a network architecture to incorporate all steps of stereo matching. The network consists of three parts. The first part calculates the multi-scale shared features. The second part performs matching cost calculation, matching cost aggregation and disparity calculation to estimate the initial disparity using shared features. The initial disparity and the shared features are used to calculate the feature constancy that measures correctness of the correspondence between two input images. The initial disparity and the feature constancy are then fed to a subnetwork to refine the initial disparity. The proposed method has been evaluated on the Scene Flow and KITTI datasets. It achieves the state-of-the-art performance on the KITTI 2012 and KITTI 2015 benchmarks while maintaining a very fast running time.
Despite the remarkable progress in face recognition related technologies, reliably recognizing faces across ages still remains a big challenge. The appearance of a human face changes substantially over time, resulting in significant intra-class variations. As opposed to current techniques for age-invariant face recognition, which either directly extract age-invariant features for recognition, or first synthesize a face that matches target age before feature extraction, we argue that it is more desirable to perform both tasks jointly so that they can leverage each other. To this end, we propose a deep Age-Invariant Model (AIM) for face recognition in the wild with three distinct novelties. First, AIM presents a novel unified deep architecture jointly performing cross-age face synthesis and recognition in a mutual boosting way. Second, AIM achieves continuous face rejuvenation/aging with remarkable photorealistic and identitypreserving properties, avoiding the requirement of paired data and the true age of testing samples. Third, we develop effective and novel training strategies for end-to-end learning the whole deep architecture, which generates powerful age-invariant face representations explicitly disentangled from the age variation. Moreover, we propose a new large-scale Cross-Age Face Recognition (CAFR) benchmark dataset to facilitate existing efforts and push the frontiers of age-invariant face recognition research. Extensive experiments on both our CAFR and several other cross-age datasets (MORPH, CACD and FG-NET) demonstrate the superiority of the proposed AIM model over the state-of-thearts. Benchmarking our model on one of the most popular unconstrained face recognition datasets IJB-C additionally verifies the promising generalizability of AIM in recogniz- *
In this paper, we study the challenging unconstrained set-based face recognition problem where each subject face is instantiated by a set of media (images and videos) instead of a single image. Naively aggregating information from all the media within a set would suffer from the large intraset variance caused by heterogeneous factors (e.g., varying media modalities, poses and illuminations) and fail to learn discriminative face representations. A novel Multi-Prototype Network (MPNet) model is thus proposed to learn multiple prototype face representations adaptively from the media sets. Each learned prototype is representative for the subject face under certain condition in terms of pose, illumination and media modality. Instead of handcrafting the set partition for prototype learning, MPNet introduces a Dense SubGraph (DSG) learning sub-net that implicitly untangles inconsistent media and learns a number of representative prototypes. Qualitative and quantitative experiments clearly demonstrate superiority of the proposed model over state-of-the-arts.
Here, a deep learning (DL) algorithm based on deep neural networks is proposed and employed to predict the chiroptical response of two-dimensional (2D) chiral metamaterials. Specifically, these 2D metamaterials contain nine types of left-handed nanostructure arrays, including U-like, T-like, and I-like shapes. Both the traditional rigorous coupled wave analysis (RCWA) method and DL approach are utilized to study the circular dichroism (CD) in higher-order diffraction beams. One common feature of these chiral metamaterials is that they all exhibit the weakest intensity but the strongest CD response in the third-order diffracted beams. Our work suggests that the DL model can predict CD performance of a 2D chiral nanostructure with a computational speed that is four orders of magnitude faster than RCWA but preserves high accuracy. The DL model introduced in this work shows great potentials in exploring various chiroptical interactions in metamaterials and accelerating the design of hypersensitive photonic devices.
Metamaterials play an important role in the modulation of amplitude and group delay in the terahertz (THz) regime on account of their optical properties, which are rare in natural materials. Here an ultrafast anisotropic switch of the plasmon-induced transparency (PIT) effect is experimentally and numerically demonstrated by metamaterial devices composed of two pairs of planar split-ring resonators and a pair of closed-ring resonators. By integration with a germanium (Ge) film, a recovery time of 3 ps and a decay constant of 785 fs are realized in the metadevice. Stimulated by the exterior optical pump, the PIT windows at different frequencies are switched off with an excellent property of slow light for vertical and horizontal THz polarizations, realizing an astonishing modulation depth as high as 99.06%. This work provides a new platform for ultrafast anisotropic metadevices tunable for amplitude and group delay.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.