Lijuan Duan

Visual saliency detection by spatially weighted dissimilarity

In this paper, a new visual saliency detection method is proposed based on the spatially weighted dissimilarity. We measured the saliency by integrating three elements as follows: the dissimilarities between image patches, which were evaluated in the reduced dimensional space, the spatial distance between image patches and the central bias. The dissimilarities were inversely weighted based on the corresponding spatial distance. A weighting mechanism, indicating a bias for human fixations to the center of the image, was employed. The principal component analysis (PCA) was the dimension reducing method used in our system. We extracted the principal components (PCs) by sampling the patches from the current image. Our method was compared with four saliency detection approaches using three image datasets. Experimental results show that our method outperforms current state-of-the-art methods on predicting human fixations.

Detecting adult image using multiple features

This paper presents a new method to detect naked people in an image using multiple features. The skin color model is firstly used to detect naked skin areas roughly. The Sobel edge operator and Gabor filter are used to weed those that are not really human skin pixels. Images that have many naked skin areas are thought maybe to have naked people. The color coherence vector and color histogram of these images are calculated and the SVM (support vector machine) is used to determine which of these images contain images of naked people.

A Voting Optimized Strategy Based on ELM for Improving Classification of Motor Imagery BCI Data

This paper presents an approach to classifying electroencephalogram (EEG) signals for brain–computer interfaces (BCI). To eliminate redundancy in high-dimensional EEG signals and reduce the coupling among different classes of EEG signals, we use principle component analysis and linear discriminant analysis to extract features that represent the raw signals. Next, we introduce the voting-based extreme learning machine to classify the features. Experiments performed on real-world data from the 2003 BCI competition indicate that our classification method outperforms state-of-the-art methods in speed and accuracy.

Visual Conspicuity Index: Spatial Dissimilarity, Distance, and Central Bias

We propose an image conspicuity index that combines three factors: spatial dissimilarity, spatial distance and central bias. The dissimilarity between image patches is evaluated in a reduced dimensional principal component space and is inversely weighted by the spatial separations between patches. An additional weighting mechanism is deployed that reflects the bias of human fixations towards the image center. The method is tested on three public image datasets and a video clip to evaluate its performance. The experimental results indicate highly competitive performance despite the simple definition of the proposed index. The conspicuity maps generated are more consistent with human fixations than prior state-of-the-art models when tested on color image datasets. This is demonstrated using both receiver operator characteristics (ROC) analysis and the Kullback-Leibler distance metric. The method should prove useful for such diverse image processing tasks as quality assessment, segmentation, search, or compression. The high performance and relative simplicity of the conspicuity index relative to other much more complex models suggests that it may find wide usage.

Detection of Front-View Vehicle with Occlusions Using AdaBoost

In this paper, we propose a vehicle detection method based on AdaBoost. We focus on the detection of front-view car and bus with occlusions on highway. Samples with different occlusion situations are selected into the training set. By using basic and rotated Haar-like features extracted from the samples in the set, we train an AdaBoost-based cascade vehicle detector. The performance tests on static images and short time videos show that (1)our approach detects cars more effectively than buses (2)the real-time detection of our method on video proceeds at 30 frames per second.

Motor Imagery EEG Classification Based on Kernel Hierarchical Extreme Learning Machine

As connections from the brain to an external device, Brain-Computer Interface (BCI) systems are a crucial aspect of assisted communication and control. When equipped with well-designed feature extraction and classification approaches, information can be accurately acquired from the brain using such systems. The Hierarchical Extreme Learning Machine (HELM) has been developed as an effective and accurate classification approach due to its deep structure and extreme learning mechanism. A classification system for motor imagery EEG signals is proposed based on the HELM combined with a kernel, herein called the Kernel Hierarchical Extreme Learning Machine (KHELM). Principle Component Analysis (PCA) is used to reduce the dimensionality of the data, and Linear Discriminant Analysis (LDA) is introduced to push the features away from different classes. To demonstrate the performance, the proposed system is applied to the BCI competition 2003 Dataset Ia, and the results are compared with those from state-of-the-art methods; we find that the accuracy is up to 94.54%.

A Method of Human Skin Region Detection Based on PCNN

A method of human skin region detection based on PCNN is proposed in this paper. Firstly, the input origin image is translated from RGB color space to YIQ color space, and I channel image is obtained. Secondly, we use the synchronous pulse firing mechanism of pulse coupled neural network (PCNN) to simulate the skin region detection mechanism of human eyes. Skin and non-skin regions are fired in different time. Therefore, skin regions are detected. Our comparison with other methods shows that the proposed method produces more accurate segmentation results.

Learning semantic cluster for image retrieval using association rule hypergraph partitioning

Semantic clustering is an important and challenging task for content-based image database management. This paper proposes a semantic clustering learning technique, which collects the relevance feedback image retrieval transaction and uses hypergraph to represent images correlation ship, then obtains the semantic clusters by hypergraph partitioning. Experiments show that it is efficient and simple.

Feature Extraction of Motor Imagery EEG Based on Extreme Learning Machine Auto-encoder

Feature extraction plays an important role in brain computer interface system that significantly affects the success of brain signal classification. In this paper, a feature extraction method of electroencephalographic (EEG) signals based on Extreme Learning Machine auto-encoder (ELM-AE) is applied. Firstly, the original data is classified by Extreme Learning Machine (ELM) and the number of hidden layer’s neuron with the highest accuracy is selected as the dimension of feature extraction. Then, ELM-AE’s output weight learns to represent the features of the original data. Finally, the features are classified by Support Vector Machine (SVM) classifier. Experiment result shows the efficiency of our method for both the speed of feature extraction and the accuracy of the classification for data set la, which is a typical representative of one kind of BCI competition 2003 data.

Fast interactive stereo image segmentation

The paper presents an approach to cutting out the same target object from a pair of stereo images interactively. With this approach, a user labels parts of the object and background in either of the images with strokes. The approach generates a segmentation result immediately. In case it is not satisfying, the result can be improved by interactively drawing more strokes, or using an alternative interaction way called adding corresponding points, which is first presented in this paper. The proposed segmentation approach is capable of providing feedback fast after each interaction. The fast computation is performed in the framework of graph cut. First, the labeled parts are used to learn foreground and background color models. Next, an energy function is built by formulating the similarities between unlabeled pixels and the foreground/background color models, color difference between neighbor pixels, and stereo correspondences obtained by SIFT feature matching. At last, graph cut is utilized to find the optimum of the energy function and obtain a segmentation result. Different from state-of-the-art methods, our segmentation approach formulates sparse correspondences rather than dense matches as stereo constraints in the energy function. Experimental results demonstrate that our method is faster in computation. In the meanwhile, it generates comparable results with state-of-the-art methods.