scispace - formally typeset
Search or ask a question
Author

Gu Jia

Bio: Gu Jia is an academic researcher from Shanghai University of Engineering Sciences. The author has contributed to research in topics: Feature (computer vision) & Image segmentation. The author has an hindex of 2, co-authored 6 publications receiving 17 citations.

Papers
More filters
Journal ArticleDOI
TL;DR: A novel self-supervised depth estimation network is proposed that outperforms the state-of-the-art approaches of depth estimation and uses quadtree-based Photometric loss, which calculates the averaged photometric loss in quadtree blocks instead of the pixel-wise loss.
Abstract: Depth estimation from a camera is an important task for 3D perception. Recently, without using the labeled ground truth of depth map, a self-supervised deep learning network can use relative pose to synthesize the target image from the reference image, and the photometric error between synthesized reference image and real one is used as self-supervisory signal. In this paper, we propose a novel self-supervised depth estimation network, which takes advantage of the quadtree constraint to optimize the depth estimation network. Based on the quadtree constraint, the photometric loss and depth loss of quadtree are proposed. In order to solve the problem that multiple depth values in repeated structures and uniform texture regions can cause relatively low photometric loss, we use quadtree-based photometric loss, which calculates the averaged photometric loss in quadtree blocks instead of the pixel-wise loss. For the problem of imbalanced depth distribution, we use quadtree depth loss, which constrains the depth inconsistency within quadtree blocks. The depth estimation network is composed of deep fusion module and cross-layer feature fusion module, which can better extract the feature information of RGB image and sparse keypoints depths, and makes full use of the detail information of the shallow feature map and the semantic information of the deep feature map to enrich the feature information extraction. Experimental results demonstrate that our method outperforms the state-of-the-art approaches of depth estimation.

37 citations

Journal ArticleDOI
TL;DR: This paper presents a novel CCTA image segmentation framework that combines deep learning and digital image processing algorithms to address these challenging problems and shows that the method is better than the mainstream baseline.
Abstract: The automatic segmentation of coronary artery in coronary computed tomography angiography (CCTA) image is of great significance for clinicians to evaluate patients with coronary heart disease. When a 3D image is limited by the amount of available GPU memory, reducing the resolution of 3D image will easily lead to the loss of image detail information. Taking patches of image as input cannot make full use of image context information. Image segmentation based on deep learning is difficult to recover perfect smooth edges. The use of smooth loss function may filter out some small lesions on the coronary artery. In this paper, we present a novel CCTA image segmentation framework that combines deep learning and digital image processing algorithms to address these challenging problems. We first use V-Net to process the CCTA image with lower resolution, and get the basic feature map (rough segmentation result) with the same resolution as the original CCTA image. Then, the original CCTA image is concatenated to the basic feature map and input it into the patch-based cascaded V-shaped module to obtain a accurate coronary artery segmentation image. Finally, the center points of coronary segmentation image and the basic gradient image of the original coronary image are obtained by morphological operation. The center points of coronary artery segmentation image are used as seed points, region growing is performed on the binary basic gradient image until the white contour boundary is searched, so as to obtain a coronary segmentation result with full segmentation and smooth edges. The proposed method is analyzed quantitatively and qualitatively, and the results show that the method is better than the mainstream baseline. The ablation experiment also proved the effectiveness of each module.

13 citations

Journal ArticleDOI
TL;DR: A novel global feature embedded network for better coronary arteries segmentation in 3D coronary computed tomography angiography (CTA) data is proposed, which contains semantic information and detailed features, aiming to accurately segment target with precise boundary.

12 citations

Journal ArticleDOI
TL;DR: A depth network of contour and gradient attention is proposed, which is used to complete and correct depth maps to obtain high-resolution and high-quality depth maps and significantly improves the quality of the depth maps, the localization results, and the effect of 3D reconstruction.

5 citations

Patent
18 Feb 2020
TL;DR: In this article, a man-machine interaction intelligent robot dog with four legs, a neck, a trunk, a head, a tail and a control system, the control system comprises a motion control module, a detection module, information output module, communication module, positioning navigation module and a power module and the communication module is further in bidirectional connection with a household appliance control module.
Abstract: The utility model relates to a man-machine interaction intelligent robot dog. The robot comprises four legs, a neck, a trunk, a head, a tail and a control system, the control system comprises a motioncontrol module, a detection module, an information output module, a communication module, a positioning navigation module and a power module which are in bidirectional connection with a controller, and the communication module is further in bidirectional connection with a household appliance control module. Compared with the prior art, according to the technical scheme of the utility model, the ROS building controller is adopted as the core. Based on the motion control module, the detection module, the information output module, the household electrical appliance control module, the communication module, the positioning navigation module and the power supply module, the intelligent household electrical appliance control system has the functions of multi-degree-of-freedom motion, intelligent security and protection, household electrical appliance control, environment monitoring, intelligent navigation obstacle avoidance, automatic wireless charging and man-machine interaction, and canmeet various requirements of people for household service.

2 citations


Cited by
More filters
Journal ArticleDOI
TL;DR: Wang et al. as mentioned in this paper proposed a new real-time small object detection (RSOD) algorithm based on YOLOv3, which improves the small-object detection accuracy by using feature maps of a shallower layer containing more fine-grained information for location prediction; fusing local and global features of shallow and deep feature maps in Feature Pyramid Network(FPN) to enhance the ability to extract more representative features; and improving the excitation layer in Squeeze-and-Excitation attention mechanism to adjust the feature responses of each channel more precisely.
Abstract: The prevailing applications of Unmanned Aerial Vehicles (UAVs) in transportation systems promote the development of object detection methods to collect real-time traffic information through UAVs. However, due to the small size and high density of objects from the aerial perspective, most existing algorithms are difficult to accurately process and extract informative features from the traffic images collected by UAVs. To address the challenges, this paper proposes a new real-time small object detection (RSOD) algorithm based on YOLOv3, which improves the small object detection accuracy by (i) using feature maps of a shallower layer containing more fine-grained information for location prediction; (ii) fusing local and global features of shallow and deep feature maps in Feature Pyramid Network(FPN) to enhance the ability to extract more representative features; (iii)assigning weights to output features of FPN and fusing them adaptively; and(iv) improving the excitation layer in Squeeze-and-Excitation attention mechanism to adjust the feature responses of each channel more precisely. Experimental results show that, when the input size is 608 × 608 × 3, the precision of the proposed RSOD algorithm measured by mAP@0.5 is 43.3% and 52.7% on the Visdrone-DET2018 and UAVDT datasets, which is 3.4% and 5.1% higher than those of YOLOv3, respectively.

35 citations

Journal ArticleDOI
TL;DR: In this paper, a fully automatic two-dimensional Unet model is proposed to segment the aorta and coronary arteries on CTCA images, which achieves 91.20% and 88.80% dice similarity coefficient accuracy on regions of interest 1 and 2 respectively.
Abstract: Early detection and diagnosis of coronary artery disease could reduce the risk of developing a heart attack. The coronary arteries are optimally visualised using computed tomography coronary angiography (CTCA) imaging. These images are reviewed by specialist radiologists who evaluate the coronary arteries for potential narrowing. A lack of radiologists in the UK is a constraint to timely diagnosis of coronary artery disease, particularly in the acute accident and emergency department setting. The development of automated methods by which coronary artery narrowing can be identified rapidly and accurately are therefore timely. Such complex computer based tools also need to be sufficiently computationally efficient that they can run on servers typically found in hospital settings, where graphical processing units for example are unavailable. We propose a fully automatic two-dimensional Unet model to segment the aorta and coronary arteries on CTCA images. Two models are trained to segment two regions of interest, (1) the aorta and the coronary arteries or (2) the coronary arteries alone. Our method achieves 91.20% and 88.80% dice similarity coefficient accuracy on regions of interest 1 and 2 respectively. Compared with a semi-automatic segmentation method, our model performs better when segmenting the coronary arteries alone. The performance of the proposed method is comparable to existing published two-dimensional or three-dimensional deep learning models. Importantly, the algorithmic and graphical processing unit memory efficiencies are maintained such that the model can be deployed without requiring graphical processing units, and therefore can be used in a hospital setting.

20 citations

Journal ArticleDOI
TL;DR: This paper proposes an outlier masking technique that considers the occluded or dynamic pixels as statistical outliers in the photometric error map and proposes an efficient weighted multi-scale scheme to reduce the artifacts in the predicted depth maps.
Abstract: As a flexible passive 3D sensing means, unsupervised learning of depth from monocular videos is becoming an important research topic. It utilizes the photometric errors between the target view and the synthesized views from its adjacent source views as the loss instead of the difference from the ground truth. Occlusion and scene dynamics in real-world scenes still adversely affect the learning, despite significant progress made recently. In this paper, we show that deliberately manipulating photometric errors can efficiently deal with these difficulties better. We first propose an outlier masking technique that considers the occluded or dynamic pixels as statistical outliers in the photometric error map. With the outlier masking, the network learns the depth of objects that move in the opposite direction to the camera more accurately. To the best of our knowledge, such cases have not been seriously considered in the previous works, even though they pose a high risk in applications like autonomous driving. We also propose an efficient weighted multi-scale scheme to reduce the artifacts in the predicted depth maps. Extensive experiments on the KITTI dataset and additional experiments on the Cityscapes dataset have verified the proposed approach's effectiveness on depth or ego-motion estimation. Furthermore, for the first time, we evaluate the predicted depth on the regions of dynamic objects and static background separately for both supervised and unsupervised methods. The evaluation further verifies the effectiveness of our proposed technical approach and provides some interesting observations that might inspire future research in this direction.

11 citations

Journal ArticleDOI
TL;DR: Wang et al. as discussed by the authors proposed a learned snakes model for 3D medical image segmentation, where both the initial and final surfaces are estimated using deep neural networks in end-to-end regimes.

7 citations

Journal ArticleDOI
TL;DR: This review describes the multiple steps and recent advances in a typical patient-specific simulation pipeline, including medical imaging, image processing, spatial discretization to generate computational mesh, setting up boundary conditions and solver parameters, visualization and extraction of hemodynamic factors, and statistical analysis.
Abstract: Hemodynamic factors, induced by pulsatile blood flow, play a crucial role in vascular health and diseases, such as the initiation and progression of atherosclerosis. Computational fluid dynamics, finite element analysis, and fluid-structure interaction simulations have been widely used to quantify detailed hemodynamic forces based on vascular images commonly obtained from computed tomography angiography, magnetic resonance imaging, ultrasound, and optical coherence tomography. In this review, we focus on methods for obtaining accurate hemodynamic factors that regulate the structure and function of vascular endothelial and smooth muscle cells. We describe the multiple steps and recent advances in a typical patient-specific simulation pipeline, including medical imaging, image processing, spatial discretization to generate computational mesh, setting up boundary conditions and solver parameters, visualization and extraction of hemodynamic factors, and statistical analysis. These steps have not been standardized and thus have unavoidable uncertainties that should be thoroughly evaluated. We also discuss the recent development of combining patient-specific models with machine-learning methods to obtain hemodynamic factors faster and cheaper than conventional methods. These critical advances widen the use of biomechanical simulation tools in the research and potential personalized care of vascular diseases.

7 citations