Hong Pan
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Research Subjects
Hong Pan has a total of 30 co-authors. Here are the top co-authors by number of shared publications:
Liang-Zheng Xia
2 shared publications- Ngai Fong Law
The Hong Kong Polytechnic University
2 shared publications Noor- A- Rahim
2 shared publications- Pei Li
Southeast Univ, Sch Automat, Key Lab Measurement & Control CSE, Minist Educ, Nanjing, Peoples R China
2 shared publications - Xinde Li
School of Automation, Southeast University, Nanjing, China
2 shared publications - Ah-Hwee Tan
School of Computing and Information Systems, Singapore Management University, Singapore, Singapore
1 shared publication
Latest Publications
- 1 February 2021
- Published by Elsevier in Pattern Recognition
- Vol. 110
Place perception from the fusion of different image representation
- 1 February 2021
- Published by Elsevier in Pattern Recognition
- Vol. 110
Place perception from the fusion of different image representation
- Open Access
- 5 November 2020
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 8, 208643-208657
Multi-Scale Vehicle Detection in High-Resolution Aerial Images With Context Information
Recently, unmanned aerial vehicles (UAV) are widely used in many fields due to the low cost and high flexibility. One of the most popular applications of UAV is vehicle detection in aerial images which plays an important role in traffic surveillance and urban planning. Although, many deep learning based detectors have achieved state-of-the-art (SOTA) performance in natural images, the significant variation in object scales caused by the altitude change of the UAV platform brings great challenges to these detectors for precise localization of vehicles in aerial images. To improve the detection performance for vehicles with different scales, we propose a novel detection algorithm which consists of three stages. In the first stage, to reduce the distortion of vehicles during image resizing and keep more information of aerial images, we utilize an image cropping strategy to divide the image into two patches. In the second stage, we combine the original image and two patches into a batch and detect vehicles with a Convolutional Neural Network (CNN). For feature representation in our detector, we propose Scale-specific Prediction to strengthen the multi-scale features of vehicles with context information. In the final stage, to fuse detections and suppress false alarms, we propose an Outlier-Aware Non-Maximum Suppression. Extensive experiments are conducted to demonstrate the superiority of the proposed algorithm by comparison with other SOTA solutions.
- 5 November 2020
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 8, 208643-208657
Open AccessMulti-Scale Vehicle Detection in High-Resolution Aerial Images With Context Information
Recently, unmanned aerial vehicles (UAV) are widely used in many fields due to the low cost and high flexibility. One of the most popular applications of UAV is vehicle detection in aerial images which plays an important role in traffic surveillance and urban planning. Although, many deep learning based detectors have achieved state-of-the-art (SOTA) performance in natural images, the significant variation in object scales caused by the altitude change of the UAV platform brings great challenges to these detectors for precise localization of vehicles in aerial images. To improve the detection performance for vehicles with different scales, we propose a novel detection algorithm which consists of three stages. In the first stage, to reduce the distortion of vehicles during image resizing and keep more information of aerial images, we utilize an image cropping strategy to divide the image into two patches. In the second stage, we combine the original image and two patches into a batch and detect vehicles with a Convolutional Neural Network (CNN). For feature representation in our detector, we propose Scale-specific Prediction to strengthen the multi-scale features of vehicles with context information. In the final stage, to fuse detections and suppress false alarms, we propose an Outlier-Aware Non-Maximum Suppression. Extensive experiments are conducted to demonstrate the superiority of the proposed algorithm by comparison with other SOTA solutions.
- 1 May 2020
- Published by Elsevier in Neurocomputing
- Vol. 390, 239-247
Text-based indoor place recognition with deep neural network
- 1 May 2020
- Published by Elsevier in Neurocomputing
- Vol. 390, 239-247
Text-based indoor place recognition with deep neural network
- 1 September 2016
- Published by Elsevier in Pattern Recognition Letters
- Vol. 80, 239-248
Feature representation of RGB-D images using joint spatial-depth feature pooling
- 1 September 2016
- Published by Elsevier in Pattern Recognition Letters
- Vol. 80, 239-248
Feature representation of RGB-D images using joint spatial-depth feature pooling
- 17 November 2008
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Neural Networks
- Vol. 19 (12) , 2132-2149
Efficient Object Recognition Using Boundary Representation and Wavelet Neural Network
Wavelet neural networks combine the functions of time–frequency localization from the wavelet transform and of self-studying from the neural network, which make them particularly suitable for the classification of complex patterns. In this paper, an efficient object recognition method using boundary representation and the wavelet neural network is proposed. The method employs a wavelet neural network (WNN) to characterize the singularities of the object curvature representation and to perform the object classification at the same time and in an automatic way. The local time–frequency attributes of the singularities on the object boundary are detected by making a preliminary wavelet analysis of the curvature representation. Then, the discriminative scale–translation features of the singularities are stored as the initial scale–translation parameters of the wavelet nodes in the WNN. These parameters are trained to their optimum status during the learning stage. With our approach, as opposed to matching features by convolving the signal with wavelet functions at a large number of scales, the computational burden is significantly reduced. Only a few convolutions are performed at the optimum scale–translation grids during the classification, which makes it suitable for real-time recognition tasks. Compared with the artificial-neural-network-based approaches preceded by wavelet filter banks with fixed scale–translation parameters, the support vector machine (SVM) using traditional Fourier descriptors and$K$-nearest-neighbor ($K$-NN) classifier based on the state-of-the-art shape descriptors, our scheme demonstrates superior and stable discrimination performance under various noisy and affine conditions.
- 17 November 2008
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Neural Networks
- Vol. 19 (12) , 2132-2149
Efficient Object Recognition Using Boundary Representation and Wavelet Neural Network
Wavelet neural networks combine the functions of time–frequency localization from the wavelet transform and of self-studying from the neural network, which make them particularly suitable for the classification of complex patterns. In this paper, an efficient object recognition method using boundary representation and the wavelet neural network is proposed. The method employs a wavelet neural network (WNN) to characterize the singularities of the object curvature representation and to perform the object classification at the same time and in an automatic way. The local time–frequency attributes of the singularities on the object boundary are detected by making a preliminary wavelet analysis of the curvature representation. Then, the discriminative scale–translation features of the singularities are stored as the initial scale–translation parameters of the wavelet nodes in the WNN. These parameters are trained to their optimum status during the learning stage. With our approach, as opposed to matching features by convolving the signal with wavelet functions at a large number of scales, the computational burden is significantly reduced. Only a few convolutions are performed at the optimum scale–translation grids during the classification, which makes it suitable for real-time recognition tasks. Compared with the artificial-neural-network-based approaches preceded by wavelet filter banks with fixed scale–translation parameters, the support vector machine (SVM) using traditional Fourier descriptors and$K$-nearest-neighbor ($K$-NN) classifier based on the state-of-the-art shape descriptors, our scheme demonstrates superior and stable discrimination performance under various noisy and affine conditions.
- 1 June 2008
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- No. 21614393,p. 1959-1963
Efficient and low-complexity image coding with the lifting scheme and modified SPIHT
In this paper, we propose an efficient and low complexity image coding algorithm based on the lifting wavelet transform and listless modified SPIHT (LWT-LMSPIHT). LWT-LMSPIHT jointly considers the advantages of progressive transmission and spatial scalability that were not fully provided by the SPIHT algorithm, thus it outperforms the SPIHT at low bit rates coding. The coding efficiency of LWT-LMSPIHT comes from three aspects. The lifting scheme lowers the number of arithmetic operations of the wavelet transform. Moreover, a significance reordering of the modified SPIHT ensures that it codes more significant information earlier in the bit stream belonging to the lower frequency bands than SPIHT to better exploit the energy compaction of the wavelet coefficients. Finally, a listless structure further reduces the amount of memory and improves the speed of compression by more than 47% for a 512 times 512 image, as compared with the SPIHT algorithm.
- 1 June 2008
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- No. 21614393,p. 1959-1963
Efficient and low-complexity image coding with the lifting scheme and modified SPIHT
In this paper, we propose an efficient and low complexity image coding algorithm based on the lifting wavelet transform and listless modified SPIHT (LWT-LMSPIHT). LWT-LMSPIHT jointly considers the advantages of progressive transmission and spatial scalability that were not fully provided by the SPIHT algorithm, thus it outperforms the SPIHT at low bit rates coding. The coding efficiency of LWT-LMSPIHT comes from three aspects. The lifting scheme lowers the number of arithmetic operations of the wavelet transform. Moreover, a significance reordering of the modified SPIHT ensures that it codes more significant information earlier in the bit stream belonging to the lower frequency bands than SPIHT to better exploit the energy compaction of the wavelet coefficients. Finally, a listless structure further reduces the amount of memory and improves the speed of compression by more than 47% for a 512 times 512 image, as compared with the SPIHT algorithm.
- 1 March 2008
- Published by Institute of Electrical and Electronics Engineers (IEEE) in International Conference on Acoustics, Speech, and Signal Processing (ICASSP)
- No. 15206149,p. 1965-1968
Wavelet neural network for 2D object classification
In this paper, a wavelet neural network (WNN)-based approach for invariant 2D object classification is proposed. The method employs the WNN characterizing the singularities of the object curvature representation and performing the classification at the same time and in an automatic way. The discriminative time-frequency attributes of the singularities on the object boundary are firstly captured by the continuous wavelet transform (CWT) and then stored by the WNN as its initial scale-translation parameters. These parameters are trained to the optimum status during the learning stage. Thus, only a few convolutions at the optimum scale-translation grids are involved during the classification, which makes our method suitable for real-time recognition tasks. Compared with the artificial neural network (ANN)-based approach preceded by a wavelet filter bank with fixed scale-translation parameters as well as the traditional methods like Fourier descriptors and moment invariants, our scheme demonstrates the best discrimination performance under various noisy and affine conditions.
- 1 March 2008
- Published by Institute of Electrical and Electronics Engineers (IEEE) in International Conference on Acoustics, Speech, and Signal Processing (ICASSP)
- No. 15206149,p. 1965-1968
Wavelet neural network for 2D object classification
In this paper, a wavelet neural network (WNN)-based approach for invariant 2D object classification is proposed. The method employs the WNN characterizing the singularities of the object curvature representation and performing the classification at the same time and in an automatic way. The discriminative time-frequency attributes of the singularities on the object boundary are firstly captured by the continuous wavelet transform (CWT) and then stored by the WNN as its initial scale-translation parameters. These parameters are trained to the optimum status during the learning stage. Thus, only a few convolutions at the optimum scale-translation grids are involved during the classification, which makes our method suitable for real-time recognition tasks. Compared with the artificial neural network (ANN)-based approach preceded by a wavelet filter bank with fixed scale-translation parameters as well as the traditional methods like Fourier descriptors and moment invariants, our scheme demonstrates the best discrimination performance under various noisy and affine conditions.
- 4 December 2007
- Published by Institution of Engineering and Technology (IET) in IET Image Processing
- Vol. 1 (4) , 353-362
Lossless image compression using binary wavelet transform
A binary wavelet transform (BWT) has several distinct advantages over a real wavelet transform when applied to binary data. No quantisation distortion is introduced and the transform is completely invertible. Since all the operations involved are modulo-2 arithmetic, it is extremely fast. The outstanding qualities of the BWT make it suitable for binary image-processing applications. The BWT, originally designed for binary images, is extended to the lossless compression of grey-level images. An in-place implementation structure of the BWT is explored. Then, a simple embedded lossless BWT-based image-coding algorithm called progressive partitioning binary wavelet-tree coder (PPBWC) is proposed. The proposed algorithm is simple in concept and implementation, but achieves promising lossless compression efficiency as compared with the conventional bitplane scanning methods. Small alphabets in the arithmetic coding, non-causal adaptive context modelling and source division are the major factors that contribute to the gain of compression efficiency of the PPBWC. Experimental results show that the PPBWC outperforms most of other embedded coders in terms of coding efficiency.
- 4 December 2007
- Published by Institution of Engineering and Technology (IET) in IET Image Processing
- Vol. 1 (4) , 353-362
Lossless image compression using binary wavelet transform
A binary wavelet transform (BWT) has several distinct advantages over a real wavelet transform when applied to binary data. No quantisation distortion is introduced and the transform is completely invertible. Since all the operations involved are modulo-2 arithmetic, it is extremely fast. The outstanding qualities of the BWT make it suitable for binary image-processing applications. The BWT, originally designed for binary images, is extended to the lossless compression of grey-level images. An in-place implementation structure of the BWT is explored. Then, a simple embedded lossless BWT-based image-coding algorithm called progressive partitioning binary wavelet-tree coder (PPBWC) is proposed. The proposed algorithm is simple in concept and implementation, but achieves promising lossless compression efficiency as compared with the conventional bitplane scanning methods. Small alphabets in the arithmetic coding, non-causal adaptive context modelling and source division are the major factors that contribute to the gain of compression efficiency of the PPBWC. Experimental results show that the PPBWC outperforms most of other embedded coders in terms of coding efficiency.
- 1 November 2007
- Published by Elsevier in Clinical Neuroscience Research
- Vol. 6 (6) , 391-398
A network approach to fMRI condition-dependent cognitive activation studies as applied to understanding sex differences
- 1 November 2007
- Published by Elsevier in Clinical Neuroscience Research
- Vol. 6 (6) , 391-398
A network approach to fMRI condition-dependent cognitive activation studies as applied to understanding sex differences
- 1 April 2005
- Published by Elsevier in Neural Networks
- Vol. 18 (3) , 297-306
Predictive neural networks for gene expression data analysis
- 1 April 2005
- Published by Elsevier in Neural Networks
- Vol. 18 (3) , 297-306
Predictive neural networks for gene expression data analysis