Hadamard ghost imaging with a small amount of mask plates based on a spread spectrum.

Ghost imaging techniques using low-cost bucket detectors have unrivaled advantages for some wavebands where plane array detectors are not available or where focusing is difficult. In these bands, fine mask plates are the key to implementing high-resolution and quality ghost imaging. However, manufacturing a large number of mask plates is necessary but undoubtedly expensive in traditional Hadamard ghost imaging (HGI). Inspired by the spread spectrum technology, Hadamard ghost imaging based on spread spectrum (HGI-SS) is proposed, in which only two sets of a small number of mask plates are needed to accomplish Nyquist sampling for the object. Their numbers are equal to the lateral pixel resolution and the vertical pixel resolution of the object, respectively. Optical experiments verify the effectiveness of the scheme. For ghost imaging with a resolution requirement of 128 × 128 pixels, HGI-SS needs to prepare only 256 mask plates, while the traditional HGI needs to prepare 16,384 mask plates. HGI-SS may be helpful to expand the pixel resolution of imaging at a relatively low cost of mask plates.

Digging Into Uncertainty-Based Pseudo-Label for Robust Stereo Matching.

Due to the domain differences and unbalanced disparity distribution across multiple datasets, current stereo matching approaches are commonly limited to a specific dataset and generalize poorly to others. Such domain shift issue is usually addressed by substantial adaptation on costly target-domain ground-truth data, which cannot be easily obtained in practical settings. In this paper, we propose to dig into uncertainty estimation for robust stereo matching. Specifically, to balance the disparity distribution, we employ a pixel-level uncertainty estimation to adaptively adjust the next stage disparity searching space, in this way driving the network progressively prune out the space of unlikely correspondences. Then, to solve the limited ground truth data, an uncertainty-based pseudo-label is proposed to adapt the pre-trained model to the new domain, where pixel-level and area-level uncertainty estimation are proposed to filter out the high-uncertainty pixels of predicted disparity maps and generate sparse while reliable pseudo-labels to align the domain gap. Experimentally, our method shows strong cross-domain, adapt, and joint generalization and obtains 1st place on the stereo task of Robust Vision Challenge 2020. Additionally, our uncertainty-based pseudo-labels can be extended to train monocular depth estimation networks in an unsupervised way and even achieves comparable performance with the supervised methods.

TUSR-Net: Triple Unfolding Single Image Dehazing with Self-Regularization and Dual Feature to Pixel Attention.

Single image dehazing is a challenging and illposed problem due to severe information degeneration of images captured in hazy conditions. Remarkable progresses have been achieved by deep-learning based image dehazing methods, where residual learning is commonly used to separate the hazy image into clear and haze components. However, the nature of low similarity between haze and clear components is commonly neglected, while the lack of constraint of contrastive peculiarity between the two components always restricts the performance of these approaches. To deal with these problems, we propose an end-to-end self-regularized network (TUSR-Net) which exploits the contrastive peculiarity of different components of the hazy image, i.e, self-regularization (SR). In specific, the hazy image is separated into clear and hazy components and constraint between different image components, i.e., self-regularization, is leveraged to pull the recovered clear image closer to groundtruth, which largely promotes the performance of image dehazing. Meanwhile, an effective triple unfolding framework combined with dual feature to pixel attention is proposed to intensify and fuse the intermediate information in feature, channel and pixel levels, respectively, thus features with better representational ability can be obtained. Our TUSR-Net achieves better trade-off between performance and parameter size with weight-sharing strategy and is much more flexible. Experiments on various benchmarking datasets demonstrate the superiority of our TUSR-Net over state-of-the-art single image dehazing methods.

DiT-SLAM: Real-Time Dense Visual-Inertial SLAM with Implicit Depth Representation and Tightly-Coupled Graph Optimization.

Recently, generating dense maps in real-time has become a hot research topic in the mobile robotics community, since dense maps can provide more informative and continuous features compared with sparse maps. Implicit depth representation (e.g., the depth code) derived from deep neural networks has been employed in the visual-only or visual-inertial simultaneous localization and mapping (SLAM) systems, which achieve promising performances on both camera motion and local dense geometry estimations from monocular images. However, the existing visual-inertial SLAM systems combined with depth codes are either built on a filter-based SLAM framework, which can only update poses and maps in a relatively small local time window, or based on a loosely-coupled framework, while the prior geometric constraints from the depth estimation network have not been employed for boosting the state estimation. To well address these drawbacks, we propose DiT-SLAM, a novel real-time Dense visual-inertial SLAM with implicit depth representation and Tightly-coupled graph optimization. Most importantly, the poses, sparse maps, and low-dimensional depth codes are optimized with the tightly-coupled graph by considering the visual, inertial, and depth residuals simultaneously. Meanwhile, we propose a light-weight monocular depth estimation and completion network, which is combined with attention mechanisms and the conditional variational auto-encoder (CVAE) to predict the uncertainty-aware dense depth maps from more low-dimensional codes. Furthermore, a robust point sampling strategy introducing the spatial distribution of 2D feature points is also proposed to provide geometric constraints in the tightly-coupled optimization, especially for textureless or featureless cases in indoor environments. We evaluate our system on open benchmarks. The proposed methods achieve better performances on both the dense depth estimation and the trajectory estimation compared to the baseline and other systems.

Deep Learning Enabled Diagnosis of Children's ADHD Based on the Big Data of Video Screen Long-Range EEG.

Attention-deficit hyperactivity disorder (ADHD) is a common neurodevelopmental disorder in children. At the same time, ADHD is prone to coexist with other mental disorders, so the diagnosis of ADHD in children is very important. Electroencephalogram (EEG) is the sum of the electrical activity of local neurons recorded from the extracranial scalp or intracranial. At present, there are two main methods of long-range EEG monitoring commonly used in clinical practice: one is ambulatory EEG monitoring, and the other is long-range video EEG monitoring. The purpose of this study is to summarize the brain electrical activity and clinical characteristics of children with ADHD through the video long-range computer graphics data of children with ADHD and to explore the clinical significance of video long-range EEG in the diagnosis of children with ADHD. For a more effective analysis, this study further processed the video data of long-range computer graphics of children with ADHD and constructed several neural network algorithm models based on deep learning, mainly including fully connected neural network models and two-dimensional convolutional neural networks. Model and long- and short-term memory neural network model. By comparing the recognition effects of these several algorithms, find the appropriate recognition algorithm to improve the accuracy and then establish a recognition method for the diagnosis of children's ADHD based on deep learning long-range EEG big data. Finally, it is concluded that long-term video EEG can analyze the EEG relationship of children with ADHD and provide a diagnostic basis for the diagnosis of ADHD.

Color computational ghost imaging based on a generative adversarial network.

A novel, to the best of our knowledge, color computational ghost imaging scheme is presented for the reconstruction of a color object image, which greatly simplifies the experimental setup and shortens the acquisition time. Compared to conventional schemes, it only adopts one digital light projector to project color speckles and one single-pixel detector to receive the light intensity, instead of utilizing three monochromatic paths separately and synthesizing the three branch results. Severe noise and color distortion, which are common in ghost imaging, can be removed by the utilization of a generative adversarial network, because it has advantages in restoring the image's texture details and generating the image's match to a human's subjective feelings over other generative models in deep learning. The final results can perform consistently better visual quality with more realistic and natural textures, even at the low sampling rate of 0.05.

MLDA-Net: Multi-Level Dual Attention-Based Network for Self-Supervised Monocular Depth Estimation.

The success of supervised learning-based single image depth estimation methods critically depends on the availability of large-scale dense per-pixel depth annotations, which requires both laborious and expensive annotation process. Therefore, the self-supervised methods are much desirable, which attract significant attention recently. However, depth maps predicted by existing self-supervised methods tend to be blurry with many depth details lost. To overcome these limitations, we propose a novel framework, named MLDA-Net, to obtain per-pixel depth maps with shaper boundaries and richer depth details. Our first innovation is a multi-level feature extraction (MLFE) strategy which can learn rich hierarchical representation. Then, a dual-attention strategy, combining global attention and structure attention, is proposed to intensify the obtained features both globally and locally, resulting in improved depth maps with sharper boundaries. Finally, a reweighted loss strategy based on multi-level outputs is proposed to conduct effective supervision for self-supervised depth estimation. Experimental results demonstrate that our MLDA-Net framework achieves state-of-the-art depth prediction results on the KITTI benchmark for self-supervised monocular depth estimation with different input modes and training modes. Extensive experiments on other benchmark datasets further confirm the superiority of our proposed approach.

The ApolloScape Open Dataset for Autonomous Driving and Its Application.

Autonomous driving has attracted tremendous attention especially in the past few years. The key techniques for a self-driving car include solving tasks like 3D map construction, self-localization, parsing the driving road and understanding objects, which enable vehicles to reason and act. However, large scale data set for training and system evaluation is still a bottleneck for developing robust perception models. In this paper, we present the ApolloScape dataset [1] and its applications for autonomous driving. Compared with existing public datasets from real scenes, e.g., KITTI [2] or Cityscapes [3] , ApolloScape contains much large and richer labelling including holistic semantic dense point cloud for each site, stereo, per-pixel semantic labelling, lanemark labelling, instance segmentation, 3D car instance, high accurate location for every frame in various driving videos from multiple sites, cities and daytimes. For each task, it contains at lease 15x larger amount of images than SOTA datasets. To label such a complete dataset, we develop various tools and algorithms specified for each task to accelerate the labelling process, such as joint 3D-2D segment labeling, active labelling in videos etc. Depend on ApolloScape, we are able to develop algorithms jointly consider the learning and inference of multiple tasks. In this paper, we provide a sensor fusion scheme integrating camera videos, consumer-grade motion sensors (GPS/IMU), and a 3D semantic map in order to achieve robust self-localization and semantic segmentation for autonomous driving. We show that practically, sensor fusion and joint learning of multiple tasks are beneficial to achieve a more robust and accurate system. We expect our dataset and proposed relevant algorithms can support and motivate researchers for further development of multi-sensor fusion and multi-task learning in the field of computer vision.

Treadmill exercise promotes angiogenesis in the ischemic penumbra of rat brains through caveolin-1/VEGF signaling pathways.

The purpose of this study was to investigate the role of caveolin-1 in treadmill-exercise-induced angiogenesis in the ischemic penumbra of rat brains, and whether caveolin-1 changes correlated with reduced brain injury induced by treadmill exercise, in rats after cerebral ischemia. Rats were randomized into five groups: sham-operated (S, n=7), model (M, n=36), exercise and model (EM, n=36), inhibitor and model (IM, n=36), and inhibitor, exercise, and model (IEM, n=36). Rats in the model groups underwent middle cerebral artery occlusion (MCAO). Rats in the inhibitor groups received an IP injection of the caveolin-1 inhibitor, daidzein (0.4 mg/kg), every 24 h following reperfusion. Rats were killed at 7 or 28 days after the operation. The exercise group showed better neurological recovery and smaller infarction volumes compared with the non-exercise group. Correspondingly, significant increases of caveolin-1 and vascular endothelial growth factor (VEGF) protein expression were observed compared with the non-exercise group. Additionally, the number of Flk-1/CD34 double-positive cells towards the ischemic penumbra was increased in the exercise group. Furthermore, the induction of VEGF protein, microvessel density, decrease of infarct volumes and neurological recovery was significantly inhibited by daidzein. This study indicates that treadmill exercise reduces brain injury in stroke. Our findings suggest that the caveolin-1 pathway is involved in the regulation of VEGF in association with promoted angiogenesis in the ischemic penumbra of rat brains after treadmill exercise. The caveolin-1/VEGF signaling pathway may be a potential target for therapeutic intervention in rats following MCAO.

Simultaneous transmission for an encrypted image and a double random-phase encryption key.

We propose a method to simultaneously transmit double random-phase encryption key and an encrypted image by making use of the fact that an acceptable decryption result can be obtained when only partial data of the encrypted image have been taken in the decryption process. First, the original image data are encoded as an encrypted image by a double random-phase encryption technique. Second, a double random-phase encryption key is encoded as an encoded key by the Rivest-Shamir-Adelman (RSA) public-key encryption algorithm. Then the amplitude of the encrypted image is modulated by the encoded key to form what we call an encoded image. Finally, the encoded image that carries both the encrypted image and the encoded key is delivered to the receiver. Based on such a method, the receiver can have an acceptable result and secure transmission can be guaranteed by the RSA cipher system.