scispace - formally typeset
Search or ask a question
Author

Guofeng Zou

Bio: Guofeng Zou is an academic researcher from Shandong University of Technology. The author has contributed to research in topics: Computer science & Facial recognition system. The author has an hindex of 6, co-authored 30 publications receiving 201 citations. Previous affiliations of Guofeng Zou include Harbin Engineering University & Chinese Academy of Sciences.

Papers
More filters
Journal ArticleDOI
TL;DR: An adaptive convolutional neural network (ACNN) is proposed, which can determine the structure of CNN without performance comparison and the experiment results of face recognition on ORL face database show that there is a better tradeoff between the consumption of training time and the recognition rate in ACNN.
Abstract: Convolutional neural network (CNN) has more and more applications in image recognition. However, the structure of CNN is often determined after a performance comparison among the CNNs with different structures, which impedes the further development of CNN. In this paper, an adaptive convolutional neural network (ACNN) is proposed, which can determine the structure of CNN without performance comparison. The final structure of ACNN is determined by automatic expansion according to performance requirement. First, the network is initialized by a one-branch structure. The system average error and recognition rate of the training samples are set to control the expansion of the structure of CNN. That is to say, the network is extended by global expansion until the system average error meets the requirement and when the system average error is satisfied, the local network is expanded until the recognition rate meets the requirement. Finally, the structure of CNN is determined automatically. Besides, the incremental learning for new samples can be achieved by adding new branches while keeping the original network unchanged. The experiment results of face recognition on ORL face database show that there is a better tradeoff between the consumption of training time and the recognition rate in ACNN.

89 citations

Journal ArticleDOI
TL;DR: Comparative results show that the BA-based tracker outperforms the other three trackers, namely, particle filter, meanshift and particle swarm optimization.

52 citations

Journal ArticleDOI
TL;DR: This survey will not only enable researchers to get a good overview of the state-of-the-art methods for RGB-D-based object recognition but also provide a reference for other multimodal machine learning applications, e.g., multimodals medical image fusion, audio-visual speech recognition, and multimedia retrieval and generation.
Abstract: Object recognition in real-world environments is one of the fundamental and key tasks in computer vision and robotics communities. With the advanced sensing technologies and low-cost depth sensors, the high-quality RGB and depth images can be recorded synchronously, and the object recognition performance can be improved by jointly exploiting them. RGB-D-based object recognition has evolved from early methods that using hand-crafted representations to the current state-of-the-art deep learning-based methods. With the undeniable success of deep learning, especially convolutional neural networks (CNNs) in the visual domain, the natural progression of deep learning research points to problems involving larger and more complex multimodal data. In this paper, we provide a comprehensive survey of recent multimodal CNNs (MMCNNs)-based approaches that have demonstrated significant improvements over previous methods. We highlight two key issues, namely, training data deficiency and multimodal fusion. In addition, we summarize and discuss the publicly available RGB-D object recognition datasets and present a comparative performance evaluation of the proposed methods on these benchmark datasets. Finally, we identify promising avenues of research in this rapidly evolving field. This survey will not only enable researchers to get a good overview of the state-of-the-art methods for RGB-D-based object recognition but also provide a reference for other multimodal machine learning applications, e.g., multimodal medical image fusion, audio-visual speech recognition, and multimedia retrieval and generation.

39 citations

Journal ArticleDOI
Mingliang Gao1, Liju Yin1, Guofeng Zou1, Haitao Li1, Wei Liu1 
TL;DR: Comparative results show that the CS-based tracker outperforms the other trackers, and the parameters’ sensitivity and adjustment of CS in the tracking system are experimentally studied.
Abstract: Cuckoo search (CS) is a new meta-heuristic optimization algorithm that is based on the obligate brood parasitic behavior of some cuckoo species in combination with the Levy flight behavior of some birds and fruit flies. It has been found to be efficient in solving global optimization problems. An application of CS is presented to solve the visual tracking problem. The relationship between optimization and visual tracking is comparatively studied and the parameters’ sensitivity and adjustment of CS in the tracking system are experimentally studied. To demonstrate the tracking ability of a CS-based tracker, a comparative study of tracking accuracy and speed of the CS-based tracker with six “state-of-art” trackers, namely, particle filter, meanshift, PSO, ensemble tracker, fragments tracker, and compressive tracker are presented. Comparative results show that the CS-based tracker outperforms the other trackers.

22 citations

Journal ArticleDOI
TL;DR: In this article, the working processes of multiple high-temperature superconducting pulsed power transformer (HTSPPT) modules are analyzed and two small experimental HTSPPTs are described in detail.
Abstract: For electromagnetic emission applications that need a very high amplitude and an appropriate pulsewidth, a pulsed power supply based on multiple high-temperature superconducting pulsed power transformer (HTSPPT) modules is verified in this paper. First, the working processes of multiple HTSPPT modules are analyzed and two small experimental HTSPPTs are described in detail. Then, simulation is carried out to show the major pulse characteristics of two different discharge methods. To verify the feasibility of two discharge methods, high-current testing is carried out with the two HTSPPTs. The results show that a higher amplitude of current pulse can be achieved using synchronous parallel discharge and a larger pulsewidth of current pulse can be achieved using asynchronous parallel discharge.

16 citations


Cited by
More filters
Posted Content
TL;DR: In this article, a spatiotemporal architecture for anomaly detection in videos including crowded scenes is proposed, which includes two main components, one for spatial feature representation, and one for learning the temporal evolution of the spatial features.
Abstract: We present an efficient method for detecting anomalies in videos. Recent applications of convolutional neural networks have shown promises of convolutional layers for object detection and recognition, especially in images. However, convolutional neural networks are supervised and require labels as learning signals. We propose a spatiotemporal architecture for anomaly detection in videos including crowded scenes. Our architecture includes two main components, one for spatial feature representation, and one for learning the temporal evolution of the spatial features. Experimental results on Avenue, Subway and UCSD benchmarks confirm that the detection accuracy of our method is comparable to state-of-the-art methods at a considerable speed of up to 140 fps.

332 citations

Journal ArticleDOI
TL;DR: A novel deep architecture based bearing diagnosis method is proposed using cognitive computing theory, which introduces the advantages of image recognition and visual perception to bearing fault diagnosis by simulating the cognition process of the cerebral cortex.

308 citations

Posted Content
TL;DR: In this article, the authors propose a new dataset and benchmark CORe50, specifically designed for continuous object recognition, and introduce baseline approaches for different continuous learning scenarios for object recognition applications.
Abstract: Continuous/Lifelong learning of high-dimensional data streams is a challenging research problem. In fact, fully retraining models each time new data become available is infeasible, due to computational and storage issues, while naive incremental strategies have been shown to suffer from catastrophic forgetting. In the context of real-world object recognition applications (e.g., robotic vision), where continuous learning is crucial, very few datasets and benchmarks are available to evaluate and compare emerging techniques. In this work we propose a new dataset and benchmark CORe50, specifically designed for continuous object recognition, and introduce baseline approaches for different continuous learning scenarios.

221 citations

Journal ArticleDOI
TL;DR: Results prove that the proposed Hybrid SCA-DE-based tracker can robustly track an arbitrary target in various challenging conditions than the other trackers and is very competitive compared to the state-of-the-art metaheuristic algorithms.

195 citations

Journal ArticleDOI
TL;DR: Simulation results prove that the bat algorithm with weighted harmonic centroid (WHCBA) strategy is superior to other algorithms and can save more energy compared to the standard LEACH protocol.

183 citations