anlong Yu

Visual–Tactile Fusion for Object Recognition

The camera provides rich visual information regarding objects and becomes one of the most mainstream sensors in the automation community. However, it is often difficult to be applicable when the objects are not visually distinguished. On the other hand, tactile sensors can be used to capture multiple object properties, such as textures, roughness, spatial features, compliance, and friction, and therefore provide another important modality for the perception. Nevertheless, effective combination of the visual and tactile modalities is still a challenging problem. In this paper, we develop a visual–tactile fusion framework for object recognition tasks. This paper uses the multivariate-time-series model to represent the tactile sequence and the covariance descriptor to characterize the image. Further, we design a joint group kernel sparse coding (JGKSC) method to tackle the intrinsically weak pairing problem in visual–tactile data samples. Finally, we develop a visual–tactile data set, composed of 18 household objects for validation. The experimental results show that considering both visual and tactile inputs is beneficial and the proposed method indeed provides an effective strategy for fusion.

An Efficient Method for Traffic Sign Recognition Based on Extreme Learning Machine

This paper proposes a computationally efficient method for traffic sign recognition (TSR). This proposed method consists of two modules: 1) extraction of histogram of oriented gradient variant (HOGv) feature and 2) a single classifier trained by extreme learning machine (ELM) algorithm. The presented HOGv feature keeps a good balance between redundancy and local details such that it can represent distinctive shapes better. The classifier is a single-hidden-layer feedforward network. Based on ELM algorithm, the connection between input and hidden layers realizes the random feature mapping while only the weights between hidden and output layers are trained. As a result, layer-by-layer tuning is not required. Meanwhile, the norm of output weights is included in the cost function. Therefore, the ELM-based classifier can achieve an optimal and generalized solution for multiclass TSR. Furthermore, it can balance the recognition accuracy and computational cost. Three datasets, including the German TSR benchmark dataset, the Belgium traffic sign classification dataset and the revised mapping and assessing the state of traffic infrastructure (revised MASTIF) dataset, are used to evaluate this proposed method. Experimental results have shown that this proposed method obtains not only high recognition accuracy but also extremely high computational efficiency in both training and recognition processes in these three datasets.

Diversified Key-Frame Selection Using Structured

In this paper, a structured L2,1 optimization model, which simultaneously characterizes the reconstruction capability and diversity, is proposed to provide a semantically meaningful representation of a short video clip acquired from digital cameras or a mobile robot. In this model, a mutual inhabitation penalty term is imposed to prevent similar samples from being selected simultaneously. The proposed model is highly flexible to incorporate different mutual inhabitation terms and the temporal redundancy in video is exploited to encourage the diversity. The constructed objective function is nonconvex and an iterative algorithm is developed to solve the optimization problem. The performance is evaluated using various video clips from YouTube and also based on practical video captured by an indoor mobile robot. The results clearly indicate that the proposed strategy helps the optimization model to achieve more diversified key frames than the other existing work method.

{L_{2,1}}

AbstractIn this paper, we try to address the joint optimization problem of the extreme learning machines corresponding to different features. The method is based on the L2,1 norm penalty, which encourages joint sparse coding. By adopting such a technology, the intrinsic relation between different features can be sufficiently preserved. To tackle the problem that the labeled samples is rare, we introduce the semi-supervised regularization term and seamlessly incorporate them into the particle filter framework to realize visual tracking. In addition, an online updating strategy is introduced which also exploits the large amount of unlabeled samples that are collected during the tracking period. Finally, the proposed tracking algorithm is compared to other state-of-the-arts on some challenging video sequences and shows promising results.

Optimization

As an important component of the driver assistance system or autonomous vehicle, traffic-sign recognition can provide drivers or vehicles with safety and alert information about the road. This paper proposes a new method for the task of traffic-sign recognition by employing extreme learning machine (ELM) whose infrastructure is a single-hidden-layer feed-forward network. This method includes two stages: One is the training stage which estimates the parameters of ELM based on training images of traffic signs; the other is the recognition stage which identifies each test image by using the trained ELM. Histogram-of-gradient descriptors are used as features in this proposed method. The German traffic sign recognition benchmark data set [1] with more than 50000 images of German road signs over 43 classes is used. Experimental results have shown that this proposed method achieves not only high recognition precision but also extremely low computational cost in terms of both training and recognition stages. An outstanding balance between recognition ratio and computational speed is obtained.

Multitask Extreme Learning Machine for Visual Tracking

In recent years, the use of human movements, especially hand gestures, serves as a motivating force for research in gesture modeling, analyzing and recognition. Hand gesture recognition provides an intelligent, natural, and convenient way of human-robot interaction (HRI). According to the way of the input of gestures, the current gesture recognition techniques can be divided into two categories: based on the vision and based on the data gloves. In order to cope with some problems existed in currently data glove. In this paper, we use a novel data glove called YoBu to collect data for gesture recognition. And we attempt to use extreme learning machine (ELM) for gesture recognition which has not yet found in the relevant application. In addition, we analyzed which features play an important role in classification and collect data of static gestures as well as establish a gesture dataset.

A novel method for traffic sign recognition based on extreme learning machine

Target tracking is one of the important tasks in computer vision. It aims to detect and track one or more particular objects in videos. The target and background may change in the process of tracking. In order to solve this problem, this paper proposes an online learning target tracking method based on extreme learning machine (ELM). First of all, we capture the target and background regions in the first few frames of video and extract the histograms of oriented gradients (HOG) features of regions into ELM. Secondly, using the method of sliding window to detect the candidate region after loading a new image. Finally, according to the tracking result, the classifier can be updated for online learning. In order to promote the detection speed, this method predicts a region in the current frame according to the target position of the previous frame. The predicted region is called the candidate region. Experiment results have shown that this proposed method not only achieves high accuracy but also can adapt to the changes of target and background.

Gesture recognition using data glove: An extreme learning machine method

This paper presents a new pruning extreme learning machine (N-PELM) algorithm which can generate a compact single-hidden-layer neural network (SLNN) by automatically pruning the number of hidden nodes while keep high accuracy. The proposed N-PELM algorithm initializes a SLNN by using extreme learning machine (ELM) algorithm given superfluous number of hidden nodes. The following part consists of two iterative processes. First, the significance of the hidden node is estimated and the insignificant node is removed. Secondly, once the node be removed, the existing hidden nodes are updated using ELM algorithm. These two processes continue until all of each hidden nodes is estimated or the number of the hidden nodes is small enough. Compared against other neural network algorithms, N-PELM algorithm has mainly three improvements. Firstly, the significance of hidden nodes is estimated in output layer such that the relevance of hidden nodes and classes can be estimated more precisely. Secondly, the pruning threshold is selected automatically from a base set of potential relevance threshold values using Akaike information criterion (AIC) such that the threshold can accommodate any data type. Thirdly, P-ELM uses Kullback-Leibler (KL) divergence and Jensen-Shannon (JS) divergence to measure the significance of hidden nodes. Experimental results have shown that the P-ELM algorithm can automatically achieve a reasonable compact network structure while keep comparable or much higher accuracy in classification and regression.It is difficult for Extreme Learning Machine (ELM) to estimate the number of hidden nodes used to match with the learning data. In this paper, a novel pruning algorithm based on sensitivity analysis is proposed for ELM. The measure to estimate the necessary number of hidden layer nodes is presented according to the defined sensitivity. When the measure is below the given threshold, the nodes with smaller sensitivities are removed from the existent network all together. Experimental results show that the proposed method can produce more compact neural network than some other existing similar algorithms.

An online learning target tracking method based on extreme learning machine

Object grasping using vision is one of the important functions of manipulators. Machine learning based methods have been proposed for grasp detection. However, due to the variety of grasps and 3D shapes of objects, how to effectively find the best grasp is still a challenging issue. Thus this paper presents an extreme learning machine (ELM) based method to cope with this issue. This proposed method consists of three successive modules, including candidate object detection, estimation of objects major orientations and grasp detection. In the first module, candidate object region is extracted based on depth information. In the second module, objects major orientations guide the directions for sliding windows. In the third module, a cascaded classifier is trained to identify the right grasp. ELM is used as the base classifier in the cascade. Histograms of oriented gradients (HOG) are used as features. Experimental results in benchmark dataset and real manipulators have shown that this proposed method outperforms other methods in terms of accuracy and computational efficiency.

A new pruning algorithm for extreme learning machine

Video-based traffic sign recognition is one of the most important task for unmanned autonomous vehicle. However, there always exists unavoidable outliers in the practical scenario. Therefore, robust prototype extraction from the noisy sample set is highly expected to help traffic sign recognition in video sequence. In this paper, we propose a novel approach for simultaneous prototype extraction and outlier isolation through collaborative sparse learning. The new model accounts for not only the reconstruction capability and the sparsity, but also the robustness. To solve the optimization problem, we adopt the Alternating Directional Method of Multiplier (ADMM) technology to design an iterative algorithm. Finally, the effectiveness of the approach is demonstrated by experiments on GTSRB dataset.