Accurate tooth segmentation from CBCT images is a crucial step for specialist to perform quantitative analysis, clinical diagnosis and operation. In this paper, we present a symmetric full convolutional network with residual block and Dense Conditional Random Field (DCRF), which can achieve accurate segmentation automatically for tooth images. The proposed method can not only strengthen feature propagation, but also boost feature reuse, which can be credited to the contracting path and the expanding path that extract and recover pixel cues sufficiently. To this end, we apply special deep bottleneck architectures (DBAs) and summation-based skip connection into our network to ensure accurate segmentation for much deeper neural network. Compared with previous methods which are based on conditional random field for original image intensity, our approach applies DCRF to the posterior probability generated by the proposed network. To avoid the interferences of noises around the tooth, we combine the pixel-level prediction capability of DCRF, which further enhance the segmentation performance. In the experiments, we verify the capabilities of our methods based on four evaluation indicators, which demonstrates the superiority of our method.

Humans have long fantasized about self-driving vehicles for the sake of luxury, style, safety, and ease. Free road space detection for collision avoidance and path planning is a vital part of autonomous driving vehicles. Despite many researchers focusing on free road space detection, it remains an open and challenging problem for real-world applications. Many studies have attempted to fuse depth and LiDAR features with visual features to improve the overall performance of free road space detection. However, there is no guideline on how such features should be fused to complement the visual features. Additionally, most of the previously proposed methods are computationally expensive and not suitable for real-life applications. The main motivation of this study is to realize a lightweight model that addresses these problems without compromising performance. As the LiDAR and visual features exist in different spaces, the proposed method attempts to learn various transformation and fusion operations from LiDAR features to complement the visual features. To validate the performance of the proposed method, we conduct comprehensive experiments on prominent benchmark datasets. The results of the experiments reveal the superior performance of the proposed model while being lightweight. LRDNet ranks third overall (with a minor difference) and second among LiDAR-based methods on the KITTI road benchmark dataset. Furthermore, the proposed model is the least computationally expensive among state-of-the-art methods and can be considered an optimal trade-off between speed and accuracy.

Visual dialog requires an agent to answer successive questions considering an image and dialog history, which is a classic vision-language task. Despite progress, there are still two key challenges: 1) parsing long or complex questions and answers and 2) dealing with the visual scene containing complicated interactions among entities. These challenges bring about the unsatisfactory consequence of current visual dialog methods. In this paper, we propose a novel Heterogeneous Knowledge Network (HKNet), which leverages textual sequence knowledge and graph knowledge to address the above issues. Specifically, the textual sequence knowledge is derived from the sentences that are retrieved from the image captions of the visual dialog dataset. The textual sequence knowledge can supplement essential common sense for parsing long or complex questions and answers. The graph knowledge is constructed via scene graph, which provides complete visual relationships for understanding the complicated interactions. These two kinds of heterogeneous knowledge complement each other and jointly improve the logical reasoning ability of the visual dialog. Extensive experimental results on two benchmark datasets: VisDial v0.9 and v1.0 demonstrate the superiority of the proposed HKNet. Ablation studies and visualization results further verify the effectiveness of our method.

The multi-UAV-enabled integrated sensing and communication (ISAC) system has emerged as a promising paradigm for jointly supporting wireless connectivity and environmental perception in internet of things (IoT) networks. However, the energy-intensive nature of UAVs’ operations poses significant challenges for the large-scale and long-duration deployments. Thus, we propose an energy-efficient trajectory design and resource allocation algorithm to minimize the weighted sum of UAVs’ communication-sensing energy and propulsion energy subject to the quality of service (QoS) requirements of all users in ISAC IoT networks. To be specific, a comprehensive system model is formulated, encompassing the UAVs’ dynamics, communication and sensing channel models, and QoS of the ground IoT users. Subsequently, this formulated non-convex optimization problem is decomposed into three subproblems: three-dimensional (3D) trajectory planning, power allocation, and users scheduling. Each subproblem is convexified using the successive convex approximation (SCA) and solved efficiently with CVX, while the global coordination is achieved via the alternating optimization. Abundant simulations are conducted to demonstrate that the proposed algorithm achieves an average 35.6% reduction in total energy consumption compared with the benchmark methods, while ensuring all IoT nodes meet the stringent QoS requirements. The proposed algorithm also exhibits the strong scalability, robustness, and low computational complexity, which can improve the performance of the UAV related consumer electronics in IoT networks.

Automatically reading pointer meters using deep learning has yielded promising results with high precision. However, existing methods ignore the interference with the camera brought by moving devices (e.g., patrol robots and drones), thus the persistent motion blur caused by the camera shake is not properly addressed. It is noteworthy that reading the pointer meter relies heavily on semantic segmentation of the scale and pointer within the meter. However, this can be challenging in extreme motion blur and diverse substation scenes. Moreover, reading various types of pointer meters and out-of-range pointer check remain tough issues. Thus, in this study, a full pipeline is proposed to solve the problems mentioned above. Firstly, Filter-Deblur-U-net (FD-U-net) is proposed to ensure accurate segmentation under motion blur. To be specific, FD-U-net is a one-stage network consisting of a deblurring module and a segmentation module. The segmentation loss supervises the optimization of deblurring module. And the proposed High Frequency Residual Attention (HFRA) in FD-U-net meticulously refines the details of motion-blurred image at the texture accumulated stage. Furthermore, the Judgement-Reading-Algorithm (JRA) is developed to complete readings of 35 types of meters. To ensure practical application, we propose the data augmentation strategy called Motion-Blur-MixUp (MB-MixUp) to maintain precise meter localization under motion blur. Additionally, we propose a method called Dark Channel Prior Dehaze Laplace (DCPD-Laplace) to determine whether the meter patch is motion-blurred. Experimental results have demonstrated the whole pipeline achieves state-of-the-art performance with average relative error and average reference error of only 1.54% and 0.48%, respectively.

Zero-shot learning (ZSL) aims to recognize image instances of unseen classes solely based on the semantic descriptions of the unseen classes. In this field, Generalized Zero-Shot Learning (GZSL) is a challenging problem in which the images of both seen and unseen classes are mixed in the testing phase of learning. Existing methods formulate GZSL as a semantic-visual correspondence problem and apply generative models such as Generative Adversarial Networks and Variational Autoencoders to solve the problem. However, these methods suffer from the bias problem since the images of unseen classes are often misclassified into seen classes. In this work, a novel model named the Dual Projective model for Zero-Shot Learning (DPZSL) is proposed using text descriptions. In order to alleviate the bias problem, we leverage two autoencoders to project the visual and semantic features into a latent space and evaluate the embeddings by a visual-semantic correspondence loss function. An additional novel classifier is also introduced to ensure the discriminability of the embedded features. Our method focuses on a more challenging inductive ZSL setting in which only the labeled data from seen classes are used in the training phase. The experimental results, obtained from two popular datasets—Caltech-UCSD Birds-200-2011 (CUB) and North America Birds (NAB)—show that the proposed DPZSL model significantly outperforms both the inductive ZSL and GZSL settings. Particularly in the GZSL setting, our model yields an improvement up to 15.2% in comparison with state-of-the-art CANZSL on datasets CUB and NAB with two splittings.