Search

The multi-view of an object can be used for 3D reconstruction. The method proposed in this paper generates the left and the top view of a target car through deep learning. The input of the method is only a front view of a 3D car and it isn't necessary for the depth of the 3D car. Firstly, a rough orthographic views of the 3D car is gotten from an information constraint network which is constructed by considering the structural relation between one view and the other two views. And then the rough orthographic views is transformed into large-pixel block rough view through the nearest interpolation, at the same time, the large-pixel blocks are also migrated to improve the quality of the rough orthographic views. Finally, the generative adversarial network with perception loss is used to enhance the large-pixel block view. In addition, the three views generated by the network can be used to synthesize a 3D point cloud shell. [ABSTRACT FROM AUTHOR]

This paper presents a self-supervised learning network called SP-Flow to generate keypoints in real-time for SLAM systems. Optical flows are employed to match the keypoints between two successive frames in the training process of SP-Flow. This approach enables the network to use datasets without manual annotations. To show the efficacy of our SP-Flow, we built an SP-Flow SLAM system by replacing ORB with SP-Flow in the ORB-SLAM2 system. The experimental results demonstrate that our SLAM system is able to achieve real-time performance and high accuracy with stereo or RGB-D images as inputs. • A self-supervised framework is used for SP-Flow network training. • The keypoints classified by the optical flow are considered when constructing the loss functions. • A keypoint extractor suitable for SLAM is proposed. • The robustness of SLAM in low-texture environments is improved. [ABSTRACT FROM AUTHOR]