Robust Distance Function Research Articles

A dynamic graph structure identification method of spatio-temporal correlation in an aluminum electrolysis cell

The dynamic correlation analysis of cell-spatial information (distributed anode current signal, DACS) is of great significance in the regional-refined control of industrial aluminum electrolysis cell. Due to the strong-dynamic spatio-temporal correlation of DACS and the complex dynamic cell noise, the existing methods are difficult to effectively obtain the spatio-temporal correlation analysis results of aluminum electrolysis cell. To solve these problems, a dynamic graph structure identification method of spatio-temporal correlation (DGSI-StC) in an aluminum electrolysis cell is proposed. The identified dynamic graph structure is used to describe the dynamic correlation analysis results of cell-spatial information. Specifically, a novel strongly robust distance function Edit Distance on Real sequences with Adaptive threshold (At-EDR) is proposed to weaken the impact of complex dynamic cell noise on the dynamic correlation analysis of cell-spatial information. Under the guidance of aluminum electrolysis process mechanism knowledge, the matrix-representation of dynamic cell noise information obtained from production data is used as the adaptive threshold of At-EDR to construct the dynamic graph structure. Then, based on the single-layer classical graph convolutional network, a framework for optimizing the dynamic graph structure is designed to capture the strong-dynamic spatio-temporal correlation of DACS. The framework obtains the optimal dynamic graph structure by optimizing the hyper-parameters. Finally, the experimental results on the public datasets show that the robustness of At-EDR is superior to existing methods. Meanwhile the experimental results on multiple sets of industrial aluminum electrolysis production datasets show that DGSI-StC improves the classification accuracy by 2.82% compared with existing classification methods.

A joint structure of multi-distance based metric learning for person re-identification

Person re-identification problem is a valuable computer vision task, which aims at matching pedestrian images of different cameras in a non-overlapping surveillance network. Existing metric learning based methods address this problem by learning a robust distance function. These methods learn a mapping subspace by forcing the distance of the positive pair much smaller than the negative pair by a strict constraint. The metric model is over-fitting to the training dataset. Due to drastic appearance variations, the handcrafted features are weak of representation for person re-identification. To address these problems, we propose a joint distance measure based approach, which learns a Mahalanobis distance and a Euclidean distance with a novel feature jointly. The novel feature is represented with a dictionary representation based method which considers pedestrian images of different camera views with the same dictionary. The joint distance combine the Mahalanobis distance based on metric learning method with the Euclidean distance based on the novel feature to measure the similarity between matching pairs. Extensive experiments are conducted on the publicly available bench marking datasets VIPeR and CUHK01. The identification results show that the proposed method reaches a good performance than the comparison methods.

EasyFlow: increasing the convergence basin of variational image matching with a feature‐based cost

Dense motion field estimation is a key computer vision problem. Many solutions have been proposed to compute small or large displacements, narrow or wide baseline stereo disparity, or non‐rigid surface registration, but a unified methodology is still lacking. The authors introduce a general framework that robustly combines direct and feature‐based matching. The feature‐based cost is built around a novel robust distance function that handles keypoints and weak features such as segments. It allows us to use putative feature matches to guide dense motion estimation out of local minima. The authors’ framework uses a robust direct data term. It is implemented with a powerful second‐order regularisation with external and self‐occlusion reasoning. Their framework achieves state‐of‐the‐art performance in several cases (standard optical flow benchmarks, wide‐baseline stereo and non‐rigid surface registration). Their framework has a modular design that customises to specific application needs.

Image Geo-Localization Based on Multiple Nearest Neighbor Feature Matching Using Generalized Graphs.

In this paper, we present a new framework for geo-locating an image utilizing a novel multiple nearest neighbor feature matching method using Generalized Minimum Clique Graphs (GMCP). First, we extract local features (e.g., SIFT) from the query image and retrieve a number of nearest neighbors for each query feature from the reference data set. Next, we apply our GMCP-based feature matching to select a single nearest neighbor for each query feature such that all matches are globally consistent. Our approach to feature matching is based on the proposition that the first nearest neighbors are not necessarily the best choices for finding correspondences in image matching. Therefore, the proposed method considers multiple reference nearest neighbors as potential matches and selects the correct ones by enforcing consistency among their global features (e.g., GIST) using GMCP. In this context, we argue that using a robust distance function for finding the similarity between the global features is essential for the cases where the query matches multiple reference images with dissimilar global features. Towards this end, we propose a robust distance function based on the Gaussian Radial Basis Function (G-RBF). We evaluated the proposed framework on a new data set of 102k street view images; the experiments show it outperforms the state of the art by 10 percent.

Robust diameter-based thickness estimation of 3D objects

We propose a robust thickness estimation approach for 3D objects based on the Shape Diameter Function (SDF). Our method first applies a modified strategy to estimate the local diameter with increased accuracy. We then compute a scale-dependent robust thickness estimate from a point cloud, constructed using this local diameter estimation and a variant of a robust distance function. The robustness of our method is benchmarked against several operations such as remeshing, geometric noise and artifacts common in triangle soups. The experimental results show a more stable local thickness estimation than the original SDF, and consistent segmentation results on defect-laden inputs.