A5100455654- Medical Image Segmentation Techniques
- Image Retrieval and Classification Techniques
- Advanced Image and Video Retrieval Techniques
- Advanced Image Processing Techniques
- Advanced Vision and Imaging
- Image and Signal Denoising Methods
- Image Processing Techniques and Applications
- Rough Sets and Fuzzy Logic
- Digital Media Forensic Detection
- Image and Object Detection Techniques
- Image Processing and 3D Reconstruction
- Advanced Neural Network Applications
- Advanced Steganography and Watermarking Techniques
- Handwritten Text Recognition Techniques
- Advanced Image Fusion Techniques
- Machine Learning and ELM
- Vehicle License Plate Recognition
- Advanced Computational Techniques and Applications
- Face and Expression Recognition
- Face recognition and analysis
- Fault Detection and Control Systems
- Anomaly Detection Techniques and Applications
- Computer Graphics and Visualization Techniques
- Nanowire Synthesis and Applications
- Image Enhancement Techniques
Ludong University
2015-2024
Tianjin University
2024
Beijing Institute of Technology
2018-2023
Shenzhen Institutes of Advanced Technology
2023
Chinese Academy of Sciences
2023
ZTE (China)
2021-2023
Zhuhai Institute of Advanced Technology
2023
Shandong Institute of Business and Technology
2019-2022
Dahua Technology (China)
2022
Nanchang Hangkong University
2011-2021
Monocular depth estimation is an ongoing challenge in computer vision. Recent progress with Transformer models has demonstrated notable advantages over conventional CNNs this area. However, there's still a gap understanding how these prioritize different regions 2D images and affect performance. To explore the differences between Transformers CNNs, we employ sparse pixel approach to contrastively analyze distinctions two. Our findings suggest that while excel handling global context...
To recover the corrupted pixels, traditional inpainting methods based on low-rank priors generally need to solve a convex optimization problem by an iterative singular value shrinkage algorithm. In this paper, we propose simple method for image using low rank approximation, which avoids time-consuming shrinkage. Specifically, if similar patches of are identified and reshaped as vectors, then patch matrix can be constructed collecting these patch-vectors. Due its columns being highly linearly...
In this paper, a novel spatial domain color image watermarking technique is proposed to rapidly and effectively protect the copyright of image. First, direct current (DC) coefficient 2D-DFT obtained in discussed, relationship between change each pixel DC Fourier transform proved. Then, used embed extract watermark by quantization technique. The novelties paper include three points: 1) without true 2D-DFT; 2) block found, and; 3) method has short running time strong robustness. experimental...
Video super-resolution has recently become one of the most important mobile-related problems due to rise video communication and streaming services. While many solutions have been proposed for this task, majority them are too computationally expensive run on portable devices with limited hardware resources. To address problem, we introduce first Mobile AI challenge, where target is develop an end-to-end deep learning-based that can achieve a real-time performance mobile GPUs. The...
Automatic defect detection is an important and challenging problem in industrial quality inspection. This paper proposes efficient method for tire assurance, which takes advantage of the feature similarity images to capture anomalies. The proposed algorithm mainly consists three steps. Firstly, local kernel regression descriptor exploited derive a set vectors inspected image. These are used evaluate dissimilarity pixels. Next, texture distortion degree each pixel estimated by weighted...
Stock price volatility forecasting is a hot topic in time series prediction research, which plays an important role reducing investment risk. However, the trend of stock not only depends on its historical trend, but also related social factors. This paper proposes hybrid time-series predictive neural network (HTPNN) that combines effection news. The features news headlines are expressed as distributed word vectors dimensionally reduced to optimize efficiency model by sparse automatic...
A deep neural network is difficult to train due a large number of unknown parameters. To increase trainable performance, we present moderate depth residual for the restoration motion blurring and noisy images. The proposed has only 10 layers, sparse feedbacks are added in middle last which called FbResNet. FbResNet fast convergence speed effective denoising performance. In addition, it can also reduce artificial Mosaic trace at seam patches, visually pleasant output results be produced from...
There are deficiencies about traditional algorithms in speed and accuracy, so a new algorithm for edge detection is proposed. Firstly the source image decomposed by wavelet lifting transform, then modulus maximum used high-frequency information Canny operator low-frequency during detection. Lifting speeds decomposition; self-adaptive processing of update operators well reserves local structural feature information; when decomposed, column row transformations reduce requirement memory...
A new aero gas turbine engine path component fault diagnosis method based on multi-hidden-layer extreme learning machine with optimized structure (OM-ELM) was proposed. OM-ELM employs quantum-behaved particle swarm optimization to automatically obtain the optimal network according both root mean square error training data set and norm of output weights. The proposed is applied handwritten recognition a diagnostic application compared basic ELM, two state-of-the-art deep algorithms: belief...
Abstract Image segmentation is a basic problem in medical image analysis and useful for disease diagnosis. However, the complexity of images makes difficult. In recent decades, fuzzy clustering algorithms have been preferred due to their simplicity efficiency. they are sensitive noise. To solve this problem, many using non-local information proposed, which perform well but inefficient. This paper proposes an improved algorithm utilizing nonlocal self-similarity low-rank prior segmentation....