Is this you? Claim this profile to correct it, add a bio and choose the work people see first.
Claim this profileWorks128 from public data
- 818
- Self-supervised Learning of Adversarial Example: Towards Good Generalizations for Deepfake Detection346
This work addresses the generalizable deepfake detection from a simple principle: a generalizable representation should be sensitive to diverse types of forgeries and synthesize augmented forgeries with a pool of forgery configurations and strengthen the “sensitivity” to the forgeries by enforcing the model to predict the forgery configuration.
- 223
- Perceiving and Modeling Density for Image Dehazing156
A novel Separable Hybrid Attention (SHA) module is proposed to perceive haze density by capturing positional-sensitive features in the orthogonal directions to achieve this goal and a density encoding matrix is proposed to model the uneven distribution of the haze explicitly.
- 132
- 117
- AIM 2020 Challenge on Efficient Super-Resolution: Methods and Results101
The AIM 2020 challenge on efficient single image super-resolution was to super-resolve an input image with a magnification factor x4 based on a set of prior examples of low and corresponding high resolution images with focus on the proposed solutions and results.
- 85
- HotRing: A Hotspot-Aware In-Memory Key-Value Store.80
This paper proposes a novel hotspot-aware KVS, named HotRing1, that is optimized for massively concurrent accesses to a small portion of items, and comprehensively adopts lock-free structures in its design, so that massively concurrent requests can better leverage multi-core architectures.
- Context-sensitive ranking64
This work presents context-sensitive ranking, a ranking framework that integrates structured queries and relevance ranking, and leverages and extends the materialized view research in OLAP to deliver algorithms and data structures that evaluate context- sensitive ranking efficiently.
- Self-Calibrated Efficient Transformer for Lightweight Super-Resolution61
This work presents a lightweight Self-Calibrated Efficient Transformer (SCET) network, which employs an efficient transformer to help the network obtain similar features over long distances and thus recover sufficient texture details.
- 59
- 47
- 45
- Enhanced Sparse Model for Blind Deblurring43
Comprehensive evaluations on public datasets and real-world images demonstrate the superiority of the proposed method against state-of-the-art methods in terms of both speed and accuracy.
- A reversible data hiding method for H.264 with Shamir’s (t, n)-threshold secret sharing43
The experimental results show that this new robust reversible data hiding algorithm can get more robustness, effectively avert intra-frame distortion drift and get good visual quality.
- Deep Richardson–Lucy Deconvolution for Low-Light Image Deblurring39
A data-driven approach to model the saturated pixels by a learned latent map is developed and performs favorably against state-of-the-art algorithms both quantitatively and qualitatively on synthetic and real-world images.
- Thunder crystal: a novel crowdsourcing-based content distribution platform36
This work implements a crowdsourcing-based content distribution system, Thunder Crystal, which utilizes agents' upload bandwidth to amplify the content distribution capacity, and observes that agents are well motivated to keep online most of the time and amplify thecontent distribution capacity by 10~20 times.
- 34
- 33
- Cross-View Hierarchy Network for Stereo Image Super-Resolution31
A cross-hierarchy information mining block (CHIMB) is designed that leverages channel attention and large kernel convolution attention to extract both global and local features from the intra-view, enabling the efficient restoration of accurate texture details.
- 31
- 29
- 26
- 24
- 23
- 23
- 22
- 20
- 18
- WatME: Towards Lossless Watermarking Through Lexical Redundancy17
This work presents theoretical analyses and empirical evidence demonstrating that X-Mark substantially enhances text generation fluency while maintaining watermark detectability and investigates watermark-ing’s impact on the emergent abilities of large language models, including zero-shot and few-shot knowledge recall, logical reasoning, and instruction following.
- 16
- 16
- 16
- 15
- Multi-Stage Degradation Homogenization for Super-Resolution of Face Images With Extreme Degradations15
- 15
- Zero-Shot Remote Sensing Image Dehazing Based on a Re-Degradation Haze Imaging Model14
A zero-shot RS image dehazing method based on a re-degradation haze imaging model, which directly restores the haze-free image from a single hazy image, which outperforms many state-of-the-art (SOTA) methods in processing uniform or slight/moderate non-uniform RS hazy images.
- 14
- A unified framework for improving the accuracy of all holistic face identification algorithms14
It is shown that Electoral College can be used as a framework that provides a significantly enhanced face identification process by improving the accuracy of all holistic algorithms.
- 14
- 13
- 13
- 13
- 13
- Parallel simulation by multi-instruction, longest-path algorithms12
The key idea is that the customer departure times are represented by longest-path distance in directed graphs instead of by the usual recursive equations, which leads to scalable algorithms with a high degree of parallelism that can be implemented on either MIMD or SIMD parallel computers.
- 11
- Noise Face Image Hallucination via Data-Driven Local Eigentransformation11
Experimental results show the effectiveness of the proposed Data-driven Local Eigentransformation DLE method for face hallucination by exploiting the local geometry structure of data manifold and learning a specified eigentransformed model for each observed image.
- Coupled Discriminant Multi-Manifold Analysis with Application to Low-Resolution Face Recognition11
Experimental results conducted on two standard face recognition databases demonstrate the superiority of the proposed coupled discriminant multi-manifold analysis (CDMMA).
- Fake View Analytics in Online Video Services11
This paper develops some algorithms and shows their effectiveness for detecting fake views, and investigates the extent of fake views with current VoD service providers.
- 11
- Comments on "Leading-One Prediction with Concurrent Position Correction"11
This report first point out and analyse an error in the implementation of the pre-encoding logic in the LOP module proposed in [1], and then presents a modification method.
- Image Denoise Based on Soft-Threshold and Edge Enhancement11
In this paper, tourists’ behaviors in Mount Emei listed in word heritage list are simulated based on TSBS, showing that TSBS can simulate the tourists' behaviors relative accurately and quickly.
- A “stereo” document representation for textual information retrieval11
It is shown that the new representation based on two or more “pictures” of each document taken from different view angles contributes to the enhanced performance of textual document retrieval by enhanced capability of extracting and capturing more individualistic features of the document.
- 11
- Every Attention Matters: An Efficient Hybrid Architecture for Long-Context Reasoning10
Benefiting from the high alignment between the training and inference engine operators, the models can undergo long-term, stable, and highly efficient optimization during the reinforcement learning phase, consistently maintaining SOTA performance across multiple challenging complex reasoning benchmarks.
- CARD: Semantic Segmentation With Efficient Class-Aware Regularized Decoder10
This work proposes a universal Class-Aware Regularization (CAR) approach to optimize the intra-class variance and inter-class distance during feature learning, motivated by the fact that humans can recognize an object by itself no matter which other objects it appears with.
- Knowledge‐dissemination channels: Analytics of stature evaluation10
A novel approach to conceptualizing and measuring stature of knowledge‐dissemination channels: the power‐impact (PI) technique, a flexible technique having 3 complementary variants, giving holistic insights about channel stature by accounting for both attraction of knowledge generators to a distribution channel and degree to which knowledge consumers choose to use a channel's knowledge content.
- 9
- 8
- 8
- Evaluating Journal Quality: Beyond “Expert” Journal Assessments in the IS Discipline8
It is suggested that a combined use of impact and power measures be employed in exercises aimed at evaluating relative statures of journals devoted to IS research, to account for both the consumption and production of IS research.
- 7
- Pairwise Macropixel Comparison Can Work at Least as Well as Advanced Holistic Algorithms for Face Recognition7
This paper shows that, by taking a few pixels together as one “big" pixel, called macropixel, and measuring the similarity of macropixels by simple Euclidean distance, a method that counts best matched macrop pixels indeed works very well for face recognition.
- 5
- 4
- 4
- Binary random systematic erasure code for RAID system4
A class of erasure codes satisfied the previous requirements, which referred as Binary Random Systematic erasure code, called BRS code for short is presented, designed as a systematic code to facilitate the store and recovery.
- 4
- 4
- A Weighted Regional Voting Based Ensemble of Multiple Classifiers for Face Recognition4
A two layer voting system called Weighted Regional Voting Based Ensemble of Multiple Classifiers (WREC), which can embed all available face recognition algorithms and is faster and more accurate than several other leading algorithms/approaches in every case.
- Displacement Template with Divide-&-Conquer Algorithm for Significantly Improving Descriptor Based Face Recognition Approaches4
A displacement template structure for improving descriptor based face recognition approaches by replacing each block in the template with its local description, where a concatenation of the local descriptions of the blocks, one from each pile, is taken to be a candidate description of the face.
- ROBUSTNESS INSTEAD OF ACCURACY SHOULD BE THE PRIMARY OBJECTIVE FOR SUBJECTIVE PATTERN RECOGNITION RESEARCH: STABILITY ANALYSIS ON MULTICANDIDATE ELECTORAL COLLEGE VERSUS DIRECT POPULAR VOTE4
Subjective pattern recognition is a class of pattern recognition problems, where the authors not only merely know a few, if any, the strategies their brains employ in making decisions in daily life but also have only limited ideas on the standards their brains use in determining the equality/inequality among the objects.
- Incomprehensible But Intelligible Human Logics: Toward a Data-Knowledge-Driven Trajectory Prediction Model3
The experimental results show that the DHTP model significantly outperforms the baseline model, showcasing enhanced accuracy and robustness across diverse driving conditions and rapidly converges to the global optimal solution, particularly in highly dynamic environments.
- A Recurrently Complicated Lightweight Network for Superresolution of Remote Sensing Images3
A recurrently complicated lightweight network (RCL-Net) for SR image recovery, through procedures of recurrent fluctuated complexity, which surpasses the advanced large SR model with less than 3% learnable parameters, compared to the state-of-the-art lightweight methods.
- 3
- 3
- A Copula Based Method for the Classification of Fish Species3
The proposed work develops a method for classification of the species of a fish given in an image, which is a sub-ordinate level classification problem, and has matched the classification accuracy of other proposed complex work for such problems.
- An Approach on Feature Selection of Cascaded Support Vector Machines with Particle Swarm Optimization Algorithm3
A novel approach is proposed which combines cascaded support vector machines optimized with PSO to enhance the classifier’s performance and Particle Swarm Optimization is a simple and easily implemented swarm intelligent algorithm, which show promising performance in optimization.
- 3
- Machine Learning-Based Estimation of foF2 and MUF(3000)F2 Using GNSS Ionospheric TEC Observations2
Three machine learning models using different algorithms to estimate the critical frequency of the F2 layer (foF2) and the maximum usable frequency of the F2 layer for a 3000 km circuit (MUF(3000)F2) based on the total electron content observed by global navigation satellite system receivers are developed.
- 2
- Robust Linear Subspace for Image Set Retrieval2
This paper attempts to take advantage of both dual linear regression and sparse coding for set-to-set based object recognition and demonstrates that the algorithm achieved best classification accuracy with state-of-the-art methods.
- 2
- A Robust Tracking Combined with Texture Feature and Background-Weighted Color Histogram2
Experimental results show that the proposed adaptive tracking combined background weight with color-texture histogram on the basis of mean shift algorithm to achieve accurate tracking in complex scenes and similar background is more efficient in dealing with complex background and occlusion.
- 1
- Endogenous security-based load balancing secure routing for UAVs with three-layer architecture1
The proposed solution combines endogenous security mechanism with routing, forming an endogenous security routing with a three-layer structure that can provide certain security and good load balancing capability.
- Cross-Modal Face Super-Resolution Based on Quasi-Siamese Domain Transfer Fusion Network1
A Quasi-Siamese Domain Transfer Fusion Network (QSDTFN) for CMFSR is proposed, whose two branches transfer two LR face modality to HR face modality by domain transfer, respectively.
- Automatic statistical chart analysis based on deep learning method1
The ASCAT model obtained better description quality evaluation scores in description quality under the METEOR criteria considering the precision and recall rate, or under the CIDEr criteria with the TF-IDF weight introduced.
- 1
- Evaluation method considering rural new energy utilization1
This paper provides a method that accounts for the subjectivity and objectivity of the evaluation object, classifies evaluation indicators, and resolves the issue of the same evaluation grade that can no longer be evaluated.
- 1
- 1
- Noisy practical facial super-resolution method via deformable constrained model with small dataset1
A novel position patch-based FSR method via latent Deformable Constrained Model (FSR-DCM), which uses latent deformable similarity as distance criteria to measure the patch similarity, which facilitates the FSR method to work in low-quality scene by emphasizing neighbor relationship and enlarging the distance difference.
- 1
- Fast stereo matching using image pyramid for lunar rover1
A novel hierarchical stereo matching that reconstructs the disparity map of the scene based on the pyramid image and more global matching (MGM) method that is highly increased and is also more accurate on lunar scenes from the obtained dense disparity maps.
- 1
- 1
- 1
- 1
- Block LBP Displacement Based Local Matching Approach for Human Face Recognition1
It is expected and shown by experiments that the Electoral College approach also applies to descriptor approaches other than LBP, which has been proved to be stable for political elections as well as general pattern recognition.
- 1
- 1
- Edge extraction method with multi-scale gradient1
A new method combined with multi-scale idea was proposed, which made the method adapted to strong or weak edge based on the local extremum instead of the global one and it is not to the noise.
- 1
- The Implementation and Evaluation of a Low-Power Clock Distribution Network Based on EPIC1
MCDE, a MCD technique based on explicitly parallel instruction computing (EPIC) architecture is designed and implemented to reduce the power of clock distribution network and a series of experiments show that it can effectively decrease the microprocessor power by 40%, compared with the original EPIC clock network microarchitecture.
- –
- –
- –
- Enhancing light field image super-resolution through Mamba-based spatial–angular correlation learning–
This paper proposes a novel approach for LF image SR that leverages Mamba, a state-space model (SSM) equipped with a selective scan mechanism, to effectively capture nonlocal spatial–angular correlation within EPIs.
- RepObE: Representation Learning-Enhanced Obfuscation Encryption Modular Semantic Task Framework–
This paper introduces, from an attacker's viewpoint, a novel framework called RepObE (Representation Learning-Enhanced Obfuscation Encryption Modular Semantic Task Framework) to secure semantic communication.
- –
- –
- An acceleration and optimization method of edge computing using filter level pruning–
The combination of deep separable convolution and filter-level pruning greatly improves the computational performance of the target detection algorithm on the edge device.
- –
- –
- –
- Research on MIDI Audio Watermarking Algorithm Based on Time Dynamics–
A method of embedding digital watermarking in both time and strength parameters by LSB method is proposed, which can increase the embedding capacity and reduce the loss of sound quality and you can do blind extraction.
- –
- –
- –
- –
- –
- –
- –
- –
- Nnovel approach to text camouflage in information security–
A novel approach to text camouflage based on multi-resolution characteristics of wavelet analysis has good hiding feature and subjective fraudulence, and the performance of robustness behaves well even when attacked by noise.
- –
Publication data from OpenAlex, with missing venues and authors filled in from Crossref; citation counts are the higher of OpenAlex and Semantic Scholar, last synced 2026-10-14. One-sentence summaries under some papers are written by Semantic Scholar’s model. Citation counts may be lower than on Google Scholar, which indexes more sources.
Report an error
Wrong papers, two people merged into one, or a profile that should not be here? Tell us and we will fix or hide it.