Anlong Ming

Combining topological and view-based features for 3D model retrieval

With the rapidly increasing of 3D models, the 3D model retrieval methods have been paid significant research attention. Most of the existing methods focus on taking advantage of one kind of feature. These methods can not achieve ideal retrieval results for different classes of 3D models. In this paper, we propose a novel 3D model retrieval algorithm by combining topological and view-based features. To preserve the topological structure of the 3D model, a multiresolutional reeb graph (MRG) is constructed according to the salient topological points. The view-based features are extracted from the images, which are rendered at each of the topological points. To preserve the spatial structure information of the images, we modify the bag-of-features (BOF) method by using the combined shell-sector model. We take the view-based features as the attribute information of the corresponding MRG nodes. The comparison between two 3D models is transformed to the problem of computing the similarity of the corresponding MRGs. Finally, we calculate the similarity between the query model and the models in the databases by adapting the earth mover distance method. Experimental results on two standard benchmarks show that our algorithm can achieve satisfactory retrieval performance.

Abandoned object detection in highway scene

Abandoned object detection in highway scene is one of the most crucial tasks in intelligent visual surveillance systems. However, few previous methods on abandoned object detection have focused on this important problem. In this paper, we present a new framework to detect the abandoned objects. In our framework, Gaussian mixture model (GMM) is used to model the background, but it is not updated every frame for keeping the abandoned objects in the foreground. To erase the noise caused by sunshine or wind, we bring an edge statistics feature based approach into the framework. Moreover, object tracking module is also integrated into the framework for a better abandoned object detection. Extensive experiments are conducted. The experimental results demonstrate that our proposed framework is not only real-time enough for practical application, but also have a very high detection accuracy.

View-based 3D model retrieval using two-level spatial structure

Recently, the view-based 3D model retrieval methods have received great research attentions. However, these methods are difficult to preserve the spatial structure of 3D models. In this paper, we propose a novel view-based 3D model retrieval method to solve this problem. Our method is based on the two-level (3D model-level and 2D image-level) spatial structure. Firstly, we extract the spatial structure circular descriptor (SSCD) images from 3D models. The SSCD images can preserve the spatial structure on the 3D model-level. Then, we modify the bag-of-features (BOF) method to extract view-based features from these SSCD images. The modified BOF method can preserve the spatial structure on the 2D image-level. Finally, we calculate the similarity between the query model and the models in the databases by adapting the earth mover distance method. Experimental results show that our method can achieve satisfactory retrieval performance for both the articulated models and the rigid models.

View-based 3D model retrieval with topological structure

With the rapidly increasing of 3D models, the view-based 3D model retrieval methods have received significant research attention. The previous view-based methods can achieve ideal retrieval result for the rigid models, but they only obtain poor retrieval result for the deformable models because they can not preserve the topological structure well. In this paper, we propose a view-based 3D model retrieval algorithm using topological structure. We extract the view-based features from the images rendered at the salient topological points. To preserve the topological structure of the 3D model, a multiresolutional reeb graph (MRG) is constructed according to the salient topological points. We take the view-based features as the attribute information of the corresponding MRG nodes. The comparison between two 3D models is transformed to compute the similarity of the corresponding MRGs. Experimental results on two standard benchmarks show that our algorithm can achieve satisfactory retrieval performance for both the deformable models and the rigid models.

Learning discriminative occlusion feature for depth ordering inference on monocular image

In this paper, a novel depth ordering inference approach is presented. Our main insight is to integrate the discriminative feature selection, occlusion feature learning and same-layer (S-L) relationship judgement into a uniform sparsity based classification objective, which cannot only supply the precise segmentation for the occlusion edge, but also reduce the solution space for the depth ordering inference efficiently. In addition, a novel triple descriptor is adopted to judge the foreground relationship, which is more discriminative than conversional local cues and can further reduce the solution space. The inference is executed by finding a valid path on a directed graph model. We validate our approach on the Cornell depth-order dataset and the NYU 2 dataset, and the convincing experimental results demonstrate the effectiveness of our approach.

EGMM: An enhanced Gaussian mixture model for detecting moving objects with intermittent stops

Moving object detection is one of the most important tasks in intelligent visual surveillance systems. Gaussian Mixture Model (GMM) has been most widely used for moving object detection, because of its robustness to variable scenes. However, to the best of our knowledge, existing GMM based methods can not detect moving objects which gradually stop and keep still state for a while. In this paper, we present an Enhanced Gaussian MixtureModel, called EGMM, to handle this problem. We integrate an Initial Gaussian Background Model (IGBM) and an extended Kalman filter based tracker with GMM, to enhance its performance. Experimental results show that our EGMM based method has a lower miss rate at the same false positives per image comparing to GMM based method for moving pedestrian detection, and it also has a higher detection rate for abandoned object detection comparing to GMM based method.

A non-rigid 3D model retrieval method based on scale-invariant heat kernel signature features

The number of non-rigid 3D models increases steadily in various areas. It is imperative to develop efficient retrieval system for 3D non-rigid models. As we know, global features fail to consistently describe the intra-class variability of non-rigid 3D models, the local features are more effective than global features for the retrieval of non-rigid 3D models. In this paper, we use Heat Kernel Signature (HKS) as the local features to represent non-rigid 3D models and further propose the retrieval method based on scale-invariant local features. Firstly, we extract key-points at multiple scales automatically. Then, the HKS local features are computed for each key-point. However, the HKS features are sensitive to scale. In order to solve this problem, we convert the scale problem into the translation problem using the diffusion Wavelets transform. To solve the translation problem, we use a kind of histogram equalization technique. Finally, we use the bipartite graph matching algorithm to compute similarity between the 3D models. Experimental results on two public benchmarks show that our method outperforms state-of-the-art methods for non-rigid 3D models retrieval.

Chinese paper-cut generated from human portrait

This paper presents a software package that renders artistic paper-cut of human portrait. It is an application software package suitable for the mobile devices, e.g., the iPad. The face fidelity is an essential factor in creating a portrait, while in the traditional algorithms, rendering portrait paper-cut is considerd as an inhomogeneous binarization process. Specifically, the binarization process is inhomogeneous for different facial areas, which always cannot correspond strictly to the grayscale, and lead to the edge being unsmoothed. In order to address this issue and make the portrait more similar to the person shown in a photograph, this paper presents an improved paper-cut algorithm. We first use a dynamic thresholding method to binarize the facial components, such as nose, eyes. Then, the principle of temple matching is employed to find a template for every facial component from a pre-prepared template library, which is closest to the real component. In addition, the software package on the iPad is not a simple migration from that on PC. Compared with the software on PC, our package has a friendly interface and is easier for operating. Experimental results show that we can get a more realistic portrait paper-cut using the package on the iPad.

Real-Time 3D Reconstruction of Traffic Scenes Under an Images-to-Model Framework

Timely extraction and intuitive display of vehicle information are two challenges for intelligent traffic surveillance. This paper presents a system for real-time 3D reconstruction of traffic scenes from video frames. We first propose a images-to-model (I2M) framework for images based 3D modeling. Under this framework, we take the traffic scene model apart into the static background model and dynamic vehicle models. Compared to the conventional multi-view stereo methods for object reconstruction, we design a scheme, which uses combination of vehicle information extraction and vehicle model retrieval, to reconstruct the moving vehicles in real-time. We further study how to integrate the retrieved vehicle models into the static background model to generate the desired traffic scene model. We also evaluate our system using three real traffic scenes, and the experimental results show that our proposed methods are effective.

Fast accurate pedestrian detection using a MPL-Boosted cascade of weak FIK-SVM classifiers

We address the problem of pedestrian detection in still images. Current pedestrian detection systems are hard to improve both speed and accuracy simultaneously. In order to achieve a balance between speed and accuracy, we propose a novel MPL-Boosted cascade of weak FIK-SVM classifiers. Our method achieves high recall while taking the speed-advantage of cascade-of-rejectors approach. Each feature in our algorithm corresponds to a 66-D HOG-LBP feature vector that describe a block. The weak classifiers we use are the separating hyper-plane computed by using a FIK-SVM. We use MPL-Boost to select features from a large set of possible blocks. The integral image and convoluted trilinear interpolation are used for rapid calculation of block feature. For a 320×240 image, the system can process 16 frames per second with sparse scan, while defeat the accuracy level of existing methods.