A5101499894- Image and Video Quality Assessment
- Advanced Image Fusion Techniques
- Visual Attention and Saliency Detection
- Video Surveillance and Tracking Methods
- Advanced Image and Video Retrieval Techniques
- Advanced Image Processing Techniques
- Image Enhancement Techniques
- Infrared Target Detection Methodologies
- Advanced Vision and Imaging
- Color Science and Applications
- Geological Studies and Exploration
- Image and Object Detection Techniques
- Advanced Measurement and Detection Methods
- Tactile and Sensory Interactions
- Advanced Sensor and Energy Harvesting Materials
- Image and Signal Denoising Methods
- Advanced Steganography and Watermarking Techniques
- Medical Image Segmentation Techniques
- Simulation and Modeling Applications
- Methane Hydrates and Related Phenomena
- Interactive and Immersive Displays
- Hydrocarbon exploration and reservoir analysis
Southwest Jiaotong University
2024
Samsung (China)
2020-2021
Beijing University of Posts and Telecommunications
2012-2016
University of Maryland, College Park
2014
Institute of Applied Technology
2009
Blind image quality assessment (BIQA) research aims to develop a perceptual model evaluate the of distorted images automatically and accurately without access non-distorted reference images. The state-of-the-art general purpose BIQA methods can be classified into two categories according types features used. first includes handcrafted which rely on statistical regularities natural These, however, are not suitable for containing text artificial graphics. second learning-based invariably...
Blind image quality assessment (BIQA) aims to develop quantitative measures automatically and accurately estimate perceptual without any prior information about the reference image. In this paper, we introduce a novel BIQA metric by structural luminance information, based on characteristics of human visual perception for distorted We extract features local binary pattern distribution. Besides, distribution normalized magnitudes is extracted represent changes in After extracting structures...
In this paper, we propose a novel “Opinion Free” (OF) No-Reference Video Quality Assessment (NR-VQA) algorithm based on frame-level unsupervised feature learning and hysteresis temporal pooling. The system consists of three components: extraction with max-min pooling, frame quality prediction Frame level features are first extracted by used to train linear Support Vector Regressor (SVR) for predicting scores frame. Frame-level then combined pooling obtain single video score. We tested the...
This paper presents an efficient no-reference blur metric where image quality is quantified by perceptual-based edge analysis. Instead of statistical measurement the whole image, it computes local statistics in vicinity detected edges at a largely reduced computational cost. Unlike existing objective metrics, proposed able to distinguish blurred cost-effective way. Evaluation shows its high prediction accuracy when applied Gaussian images. Experiments using LIVE database demonstrate that...
Document image quality assessment (DIQA) aims to predict the visual of degraded document images. Although definition "visual quality" can change based on specific applications, in this paper, we use OCR accuracy as a metric for and develop novel no-reference DIQA method high order statistics prediction. The proposed consists three steps. First, normalized local patches are extracted with regular grid comprehensive codebook is constructed by K-means clustering. Second, features softly...
Almost all present successful general purpose blind image quality assessment (BIQA) models are designed only towards singly and synthetically distorted images. However, real world images generally contain multiply types of distortions other than single type distortion. Even some state the art BIQA methods cannot work well on authentically In this paper we propose a novel effective method for to remedy shortage popular models. First, except traditional natural scene statistics (NSS) based...
Previous feature learning based blind image quality assessment (BIQA) methods invariably require large codebook or updating procedure to obtain satisfying performance. In this paper, we propose a novel general purpose BIQA method, local aggregation (LFA) model, which requires only much smaller without the need for updating. The proposed model consists of three steps. Firstly, normalized raw patches are extracted as features through regular grid and 100 codeword is constructed by K-means...
In this paper, we present an efficient no-reference image blur metric which is based on the analysis of spread edge and study human perception for varying contrast values. Our method calculates ratio significant edges global vertical respectively, final score a weighted average two ratios because giving different corresponding weights will improve prediction accuracy. Evaluation proposed shows its high accuracy when it applied to Gaussian blurred images. Experiments using LIVE TID dataset...
With the continued relevance of drug hydrates in pharmaceutical sciences, a comprehensive understanding hydrate and anhydrate forms is essential, not only through individual case studies but also from broader, systematic perspective. The Cambridge Structural Database (CSD) well-established database for crystal structures organic molecules here, structural features pharmaceutically relevant compounds forming were explored. Drug subsets generated further classified into separate sets free...
In this paper, we propose two novel Statistical Metric Fusion (SMF) methods for Image Quality Assessment (IQA) metric enhancement. First, local quality map is constructed from existing state-of-the-art IQA algorithm. After that several statistical indices are extracted map. Finally, the fused by Supervised (SMF-S) based on Support Vector Regression (SVR) and Unsupervised (SMF-U) Reciprocal Rank (RRF) to obtain final score, respectively. Experimental results largest public database TID2013...
In this paper, we develop a novel method for blind image quality assessment (BIQA) based on complete pixel level information. First, traditional rotation invariant uniform local binary pattern (LBP) histogram is extracted from grayscale as perceptual aware feature. Second, except the signs of differences, magnitudes differences in are also encoded by LBP, and joint between calculated part Finally, support vector regression (SVR) utilized to learn mapping combined feature human opinion...
The paper studies the problem of how to recognize aircraft targets staying at airport from satellite image database huge amount in an acceptable processing speed. Without adopting current method recognizing target picture by database, combines retrieval with recognition, and firstly uses technology pick out those images containing then utilizes recognition images. Some new methods or thoughts have been put forward about retrieval, segmentation for targets, a system has studied designed....