Real-world objects and environments are predominantly composed of edge features, including straight lines and curves. Such edges are crucial elements for various applications, such as CAD modeling, surface meshing, lane mapping, etc. However, existing traditional methods only prioritize lines over curves for simplicity in geometric modeling. To this end, we introduce EMAP, a new method for learning 3D edge representations with a focus on both lines and curves. Our method implicitly encodes 3D edge distance and direction in Unsigned Distance Functions (UDF) from multi-view edge maps. On top of this neural representation, we propose an edge extraction algorithm that robustly abstracts parametric 3D edges from the inferred edge points and their directions. Comprehensive evaluations demonstrate that our method achieves better 3D edge reconstruction on multiple challenging datasets. We further show that our learned UDF field enhances neural surface reconstruction by capturing more details.

We propose a robust geometry-constrained scale estimation approach for monocular visual odometry, which takes the camera height as an absolute reference. Visual odometry is an essential module for robot self-localization and autonomous navigation in unexplored environments. Scale recovery is an indispensable requirement for monocular visual odometry, since it compensates for the metric information lost by a single camera and helps to reduce the scale drift. When the camera height is considered the absolute reference, the precision of scale recovery depends on the accuracy of the road point selection and road geometric model calculation. However, most of the previous approaches solve these two problems sequentially, and their road point selection is based on the color model of the road or prior-knowledge-based fixed region. In this paper, we propose combining and iteratively solving these two problems. We adopt the geometric model, instead of the color model, of the road to select the road points. Furthermore, the selected road feature points are used to estimate the road model, which limits the road point selection. In detail, we segment our feature points with Delaunay triangulation and select road points based on the depth consistency and road model consistency. The experiments on the KITTI dataset show that our method achieves the best performance among state-of-the-art monocular visual odometry methods.

Scale recovery is one of the essential problems for monocular visual odometry. The camera height is usually used as an absolute reference to recover the scale. In this case, the precision of scale recovery depends on the accuracy of the road region detection and road geometrical model calculation. In previous works, road detection and road geometrical model calculation are solved sequentially: the road geometrical model calculation is based on the road detection and the road region detection is based on the color information. However, the color information of a road is not stable enough. In the proposed method, the estimated road geometrical model is taken into consideration to detect the road region as a feedback. Therefore, the road region detection and road geometrical model estimation can benefit each other. Delaunay Triangulation method is used to segment an input image to many triangles with the matched feature points as vertices. Every triangle region is classified as a road region or not by comparing their geometrical model with that of the road and the road geometrical model is updated online. We evaluate our visual odometry scale recovery method on the KITTI dataset and the results show that our method is achieving the best performance among all existing monocular visual odometry scale recovery methods without additional sensors.

Scale recovery is one of the central problems for monocular visual odometry. Normally, road plane and camera height are specified as reference to recover the scale. The performances of these methods depend on the plane recognition and height measurement of camera. In this work, we propose a novel method to recover the scale by incorporating the depths estimated from images using deep convolutional neural fields. Our method considers the whole environmental structure as reference rather than a specified plane. The accuracy of depth estimation contributes to the scale recovery. We improve the performance of depth estimation by considering two consecutive frames and egomotion of camera into our networks. The depth refinement and scale recovery are obtained iteratively. In this way, our method can eliminate the scale drift and improve the depth estimation simultaneously. The effectiveness of our method is verified on the KITTI dataset for both monocular visual odometry and depth estimation tasks.

Temporal alignment is an important preprocessing procedure for human action recognition. The challenge of temporal alignment problem is the temporal scale difference between human actions as well as the variability of each subject. Metric learning is the central problem of temporal alignment. This paper presents a nonlinear time alignment method with deep autoencoder. The spatio-temporal features obtained from the neural network contain the metric information for feature comparison. The effectiveness of our method is verified with k-nearest neighbor (k-NN) classifier on MSR-Action 3D and MSR-Daily Activity 3D datasets. Experimental results illustrate that the proposed method achieves superior performance to other metric based techniques.

Learning from demonstration (LID) is an effective method trying to generate trajectory from the demonstrations for the new specifications. We present a novel LID method by incorporating the trajectory feature metric term with the optimization-based motion planning method. The advantage of our method is that the feature of demonstrated trajectory is kept on the premise of generating feasible and specified trajectory. In order to maintain the characteristics of trajectory, spatiotemporal feature is chosen as metrics to be included in the cost function. Our method is verified through the simulation results of mini-jerk trajectories and trajectories from Pioneer3-AT wheeled mobile robot simulator platform. The experiments show that our method can balance the specification of task with the features of demonstrations.

Learning from demonstration requires reproduction of a movement in the new situation. We present an approach based on dynamic movement primitives (DMP) and Gaussian mixture model (GMM) to learning the movement from demonstration. The original DMP model use only one demonstration to generate the dynamical system of motion primitive. Our work extend the generalization ability by capturing the characteristic of movement from several demonstrations of the same skill. We test our method on the mini-jerk trajectories of static and moving target and on data collected from nonholonomic mobile robot simulator. These experiments show that our method can improve the generalization of the basic motion primitives which is crucial to the application of imitation learning.