Search

Interactive image segmentation aims at obtaining a segmentation mask for an image using simple user annotations. During each round of interaction, the segmentation result from the previous round serves as feedback to guide the user's annotation and provides dense prior information for the segmentation model, effectively acting as a bridge between interactions. Existing methods overlook the importance of feedback or simply concatenate it with the original input, leading to underutilization of feedback and an increase in the number of required annotations. To address this, we propose an approach called Focused and Collaborative Feedback Integration (FCFI) to fully exploit the feedback for click-based interactive image segmentation. FCFI first focuses on a local area around the new click and corrects the feedback based on the similarities of high-level features. It then alternately and collaboratively updates the feedback and deep features to integrate the feedback into the features. The efficacy and efficiency of FCFI were validated on four benchmarks, namely GrabCut, Berkeley, SBD, and DAVIS. Experimental results show that FCFI achieved new state-of-the-art performance with less computational overhead than previous methods. The source code is available at https://github.com/veizgyauzgyauz/FCFI., Accepted for publication at CVPR 2023

RGBD-based real-time dynamic 3D reconstruction suffers from inaccurate inter-frame motion estimation as errors may accumulate with online tracking. This problem is even more severe for single-view-based systems due to strong occlusions. Based on these observations, we propose OcclusionFusion, a novel method to calculate occlusion-aware 3D motion to guide the reconstruction. In our technique, the motion of visible regions is first estimated and combined with temporal information to infer the motion of the occluded regions through an LSTM-involved graph neural network. Furthermore, our method computes the confidence of the estimated motion by modeling the network output with a probabilistic model, which alleviates untrustworthy motions and enables robust tracking. Experimental results on public datasets and our own recorded data show that our technique outperforms existing single-view-based real-time methods by a large margin. With the reduction of the motion errors, the proposed technique can handle long and challenging motion sequences. Please check out the project page for sequence results: https://wenbin-lin.github.io/OcclusionFusion., Comment: Accepted by CVPR 2022. Project page: https://wenbin-lin.github.io/OcclusionFusion

One major problem in black-box adversarial attacks is the high query complexity in the hard-label attack setting, where only the top-1 predicted label is available. In this paper, we propose a novel geometric-based approach called Tangent Attack (TA), which identifies an optimal tangent point of a virtual hemisphere located on the decision boundary to reduce the distortion of the attack. Assuming the decision boundary is locally flat, we theoretically prove that the minimum $\ell_2$ distortion can be obtained by reaching the decision boundary along the tangent line passing through such tangent point in each iteration. To improve the robustness of our method, we further propose a generalized method which replaces the hemisphere with a semi-ellipsoid to adapt to curved decision boundaries. Our approach is free of pre-training. Extensive experiments conducted on the ImageNet and CIFAR-10 datasets demonstrate that our approach can consume only a small number of queries to achieve the low-magnitude distortion. The implementation source code is released online at https://github.com/machanic/TangentAttack., Accepted at NeurIPS 2021. The missing square term in Eqn.(13), as well as many other mistakes of the previous version, have been fixed in the current version

International audience; This paper proposes a geometric iteration algorithm for computing point projection and inversion on planar parametriccurves based on local biarc approximation. The iteration begins with initial estimation of the projection of theprescribed test point. For each iteration, we construct a biarc that locally approximates a segment on the originalcurve starting from the current projective point. Then we compute the projective point for the next iteration, as well asthe parameter corresponding to it, by projecting the test point onto this biarc. The iterative process terminates whenthe projective point satisfies the required precision. Examples demonstrate that our algorithm converges faster and isless dependent on the choice of the initial value compared to the traditional geometric iteration algorithms based onsingle-point approximation.

The polygonal approximation problem is a primary problem in computer graphics, pattern recognition, CAD/CAM, etc. InR 2, the cone intersection method (CIM) is one of the most efficient algorithms for approximating polygonal curves. With CIM Eu and Toussaint, by imposing an additional constraint and changing the given error criteria, resolve the three-dimensional weighted minimum number polygonal approximation problem with the parallel-strip error criterion (PS-WMN) underL 2 norm. In this paper, without any additional constraint and change of the error criteria, a CIM solution to the same problem with the line segment error criterion (LS-WMN) is presented, which is more frequently encountered than the PS-WMN is. Its time complexity isO(n 3), and the space complexity isO(n 2). An approximation algorithm is also presented, which takesO(n 2) time andO(n) space. Results of some examples are given to illustrate the efficiency of these algorithms.

International audience; Quadratic Bézier curves are important geometric entities in many applications. However, it was often ignored by the literature the fact that a single segment of a quadratic Bézier curve may fail to fit arbitrary endpoint unit tangent vectors. The purpose of this paper is to provide a solution to this problem, i.e., constructing G1 quadratic Bézier curves satisfying given endpoint (positions and arbitrary unit tangent vectors) conditions. Examples are given to illustrate the new solution and to perform comparison between the G1 quadratic Bézier cures and other curve schemes such as the composite geometric Hermite curves and the biarcs.

National audience; Through geomet ric observation , it is found that the nearest point on a curve is a tangent point between the curve and an off set curve of the other curve. Based on this observation , an off set method is presented for computing the minimum distance between two planar algebraic curves. The new method is geomet rically inst ructive , and can be used for computing the minimum distance between an algebraic curve and a parametric one on the same plane. For planar quadratic curves , the degree of the resulting univariate polynomial equation by our method is much lower than that of the equations in previous comparable methods , which may lead to lower computation complexity or higher robustness of the solutions.

International audience; We propose an adaptive geometry compression method with labels based on 4-point interpolatory subdivision schemes. It can work on digital curves of arbitrary dimensions. With the geometry compression method, a digital curve is adaptively compressed into several segments with di®erent compression levels. Each segment is a 4-point subdivision curve with a subdivision step. Labels are recorded in data compression to facilitate merging those segments in data decompression. In the meantime, we provide high-speed 4-point interpolatory subdivision curve generation methods for e±ciently decompressing the compressed data. For an arbitrary positive integer k, formulae of the number of the resultant control points of a 4-point subdivision curve after k subdivision steps are provided. Some formulae for calculating points at the kth ubdivision step are presented as well. The time complexity of the new approaches is O(n), where n is the number of the points in the given digital curve. Examples are provided as well to illustrate the efficiency of the proposed approaches.

National audience; The sufficient and necessary condition for the existence of linear Möbius transformations that can standardize the rational Bézier surfaces is given based on Möbius reparameterization theorem. To obtain the standard form of an arbitrary cubic rational Bézier surface, we then present a quadratic reparameterization algorithm to reparameterize the surface so that all the corner weights of the surface are equal to one. Examples are included to show the performance of the new method.

Submission Mar 31, 2007, revised.; International audience

Practical Intrusion Detection Systems (IDSs) based on data mining are facing two key problems, discovering intrusion knowledge from real-time network data, and automatically updating them when new intrusions appear. Most data mining algorithms work on labeled data. In order to set up basic data set for mining, huge volumes of network data need to be collected and labeled manually. In fact, it is rather difficult and impractical to label intrusions, which has been a big restrict for current IDSs and has led to limited ability of identifying all kinds of intrusion types. An improved unsupervised clustering-based intrusion model working on unlabeled training data is introduced. In this model, center of a cluster is defined and used as substitution of this cluster. Then all cluster centers are adopted to detect intrusions. Testing on data sets of KDDCUP’99, experimental results demonstrate that our method has good performance in detection rate. Furthermore, the incremental-learning method is adopted to detect those unknown-type intrusions and it decreases false positive rate.

Site de la conférence : http://www.io.tudelft.nl/caidcd2005/; In BRep geometric modeling system, the positions of end-points of the resultant intersection curves (called key points as well) are very important for many other operations, such as boolean operations. This paper provides a method to determine the positions of the end-points of the intersection curves between a cylinder and a quadric surface. Our algorithm ¯rst turns the original intersection problem into another intersec- tion problem between a set of line segments and a quadric surface, and obtains the parametric form of the intersection curves. Then, it computes all the key positions. Finally, it calculates the intersec- tion intervals and their corresponding intersection curves by mid-point judging method. The algo- rithm has been applied in the commercial software Gems6.0, and practical experiences show both ro- bustness and e±ciency of the new algorithm.

An unstructured previous termtetrahedral mesh generation algorithmnext term for 3D model with constraints is presented. To automatically generate a previous termtetrahedral meshnext term for model with constraints, an advancing front previous termalgorithmnext term is presented previous termbased on conforming constrained Delaunay tetrahedralizationnext term (CCDT). To reduce the number of visibility tests between vertices with respect to model faces as well as the computation of previous termconstrained Delaunaynext term tetrahedra, a sufficient condition for DT (previous termconstrained Delaunay tetrahedralizationnext term whose simplexes are all previous termDelaunay)next term existence is presented and utilized coupled to uniform grid and advancing front techniques in our previous termalgorithm. The meshnext term generator is robust and exhibits a linear time complexity for mechanical models with uniform density distribution.

A method for generating an interpolation closed G1 arc spline on a given closed point set is presented. For the odd case, i.e. when the number of the given points is odd, this paper disproves the traditional opinion that there is only one closed G1 arc spline interpolating the given points. In fact, the number of the resultant closed G1 arc splines fulfilling the interpolation condition for the odd case is exactly two. We provide an evaluation method based on the arc length as well such that the choice between those two arc splines is made automatically. For the even case, i.e. when the number of the given points is even, the points are automatically moved based on weight functions such that the interpolation condition for generating closed G1 arc splines is satisfied, and that the adjustment is small. And then, the G1 arc spline is constructed such that the radii of the arcs in the spline are close to each other. Examples are given to illustrate the method.

The highlight line model is a powerful tool in assessing the quality of a surface. It increases the flexibility of an interactive design environment. In this paper, a method to generate a highlight line model on an arbitrary triangular mesh is presented. Based on the highlight line model, a technique to remove local shape irregularities of a triangular mesh is then presented. The shape modification is done by solving a minimization problem and performing an iterative procedure. The new technique improves not only the shape quality of the mesh surface, but also the shape of the highlight line model. It provides an intuitive and yet suitable method for locally optimizing the shape of a triangular mesh.