Is this you? Claim this profile to correct it, add a bio and choose the work people see first.
Claim this profileAcademic lineage
Possible advisorsa guess from early papers, not confirmed
- Ting-Cheng ChangSuggested from co-authorship
Is this you? Claim this profile to confirm or dismiss it.
Works148 from public data
- Deep Fake Image Detection Based on Pairwise Learning301
This paper proposes a deep learning-based approach for detecting the fake images by using the contrastive loss and demonstrated that the proposed method significantly outperformed other state-of-the-art fake image detectors.
- 186
- Learning to Detect Fake Face Images in the Wild145
A deep forgery discriminator (DeepFD) to efficiently and effectively detect the computer-generated images and adopt contrastive loss in seeking the typical features of the synthesized images generated by different GANs and follow by concatenating a classifier to detect such computer- generated images.
- CNN-Based Joint Clustering and Representation Learning with Feature Drift Compensation for Large-Scale Image Data137
The proposed convolutional neural network to jointly solve clustering and representation learning in an iterative manner outperforms start-of-the-art clustering schemes in terms of accuracy and storage complexity on large-scale image sets containing millions of images.
- DRCT: Saving Image Super-Resolution away from Information Bottleneck120
The Dense-residual-connected Transformer (DRCT) is proposed, aimed at mitigating the loss of spatial information and stabilizing the information flow through dense-residual connections between layers, thereby unleashing the model’s potential and saving the model away from information bottleneck.
- AIM 2019 Challenge on Real-World Image Super-Resolution: Methods and Results110
The AIM 2019 challenge on real world super-resolution addresses the real world setting, where paired true high and low-resolution images are unavailable, and aims to advance the state-of-the-art and provide a standard benchmark for this newly emerging task.
- The Ninth NTIRE 2024 Efficient Super-Resolution Challenge Report91
This paper provides a comprehensive review of the NTIRE 2024 challenge, focusing on efficient single-image super-resolution (ESR) solutions and their outcomes, and gauge the state-of-the-art in efficient single-image super-resolution.
- SiGAN: Siamese Generative Adversarial Network for Identity-Preserving Face Hallucination90
A Siamese GAN to reconstruct HR faces that visually resemble their corresponding identities that significantly outperforms existing face hallucination GANs in objective face verification performance while achieving promising visual-quality reconstruction.
- 90
- NTIRE 2024 Challenge on Image Super-Resolution (×4): Methods and Results83
This paper reviews the NTIRE 2024 challenge on image super-resolution (×4), highlighting the solutions proposed and the outcomes obtained, and offers a comprehensive overview of current trends in this field.
- ADAM Challenge: Detecting Age-Related Macular Degeneration From Fundus Images83
The ensembling strategy and the incorporation of clinical domain knowledge were the key to improve the performance of the deep learning models in the Automatic Detection challenge on Age-related Macular degeneration (ADAM).
- 81
- Learned Smartphone ISP on Mobile NPUs with Deep Learning, Mobile AI 2021 Challenge: Report70
The target was to develop an end-to-end deep learning-based image signal processing (ISP) pipeline that can replace classical hand-crafted ISPs and achieve nearly real-time performance on smartphone NPUs.
- 52
- 1 st Workshop on Maritime Computer Vision (MaCVi) 2023: Challenge Results48
A new benchmark is introduced, called SeaDronesSee Object Detection v2, which extends the previous benchmark by including more classes and footage and is summarized in the report.
- 34
- Popularity Prediction of Social Media based on Multi-Modal Feature Mining32
A popularity prediction framework for social media based on multi-modal feature mining and an ensemble regression approach is presented that significantly outperforms other state-of-the-art regression models and could effectively improve the performance in the social media prediction challenge.
- Deep learning-based vehicle trajectory prediction based on generative adversarial network for autonomous driving applications31
A GAN (generative adversarial network)-based deep learning framework is presented in this paper for predicting the trajectories of surrounding vehicles of an autonomous vehicle in an RGB image sequence without explicit coordinate annotation to solve the main problem addressed.
- Social Media Prediction Based on Residual Learning and Random Forest29
A unified framework for the residual learning and random forest regression for social media prediction task and shows that the performance of the proposed method significantly outperforms other methods.
- 28
- Is Intensity Inhomogeneity Correction Useful for Classification of Breast Cancer in Sonograms Using Deep Neural Network?28
It is demonstrated that it is important to use preprocessing to highlight the image features and further give these features for deep learning models, in this way, the classification accuracy will be better to just use the original images forDeep learning.
- 27
- 27
- NTIRE 2025 Challenge on Real-World Face Restoration: Methods and Results26
This paper provides a review of the NTIRE 2025 challenge on real-world face restoration, highlighting the proposed solutions and the resulting outcomes and offers an in-depth overview of the latest trends in the field.
- 25
- NTIRE 2025 Challenge on Image Super-Resolution ($\times 4$): Methods and Results23
The NTIRE 2025 image super-resolution ($\times 4$) challenge, one of the associated competitions of the 10th NTIRE Workshop at CVPR 2025, aims to recover high-resolution images from low-resolution counterparts generated through bicubic downsampling with a $\times 4$ scaling factor.
- Ntire 2025 Image Shadow Removal Challenge Report22
This work examines the findings of the NTIRE 2025 Shadow Removal Challenge, with 17 teams successfully submitting their solutions during the final evaluation phase, and one focusing on reconstruction fidelity and the other on visual perception through a user study.
- Real-Time Compressed Sensing for Joint Hyperspectral Image Transmission and Restoration for CubeSat22
A real-time compressed sensing (RTCS) network designed to be lightweight and require only relatively few training samples for efficient and robust HSI reconstruction in the presence of the stripe effect and under noisy transmission conditions is proposed, offering new and vital capabilities for existing miniaturized satellite systems.
- Gradient Boost Tree Network based on Extensive Feature Analysis for Popularity Prediction of Social Posts22
A robust multi-modality feature mining framework for predicting SMP scores by incorporating additional identity-related features sourced from the official SMP dataset when a user's path alias is accessible, and the LightGBM and TabNet are employed to effectively capture intricate semantic relationships among different modality features and user-specific data.
- 20
- WWE-UIE: A Wavelet & White Balance Efficient Network for Underwater Image Enhancement17
WWE-UIE is proposed, a compact and efficient enhancement network that integrates three interpretable priors that achieves competitive restoration quality with substantially fewer parameters and FLOPs, enabling real-time inference on resource-limited platforms.
- 17
- 17
- Efficient-ROD17
A lightweight, computationally efficient, and effective network architecture to conquer the trade-off between computational efficiency and performance of radar object detection tasks is proposed.
- Development and validation of a prognostic nomogram to predict survival in patients with advanced pancreatic cancer receiving second‐line palliative chemotherapy17
Given that a wide variation in tumor response rates and survival times suggests heterogeneity among the patients with advanced pancreatic cancer who underwent second‐line chemotherapy, it is a challenge in clinical practice to identify patients who will receive the most benefit from L2 treatment.
- An Iterative Refinement Approach for Social Media Headline Prediction17
This study proposes a novel iterative refinement approach to predict the popularity score of the social media meta-data effectively and adopts an ensemble regressor to compensate the residues to further improve the prediction performance.
- 17
- PromptHSI: Universal Hyperspectral Image Restoration With Vision–Language Modulated Frequency Adaptation16
PromptHSI is the first universal AiO HSI restoration framework that is both prompt-guided and frequency-aware, and excels at both fine-grained detail recovery and global information restoration across diverse degradation scenarios, highlighting its significant potential for practical remote sensing applications.
- 16
- Image authentication with tampering localization based on watermark embedding in wavelet domain16
The experimental results show that the proposed SWACS and MRGSD schemes can efficiently identify different types of image tampering.
- Rethinking Relation between Model Stacking and Recurrent Neural Networks for Social Media Prediction15
A more dominant feature representation of text information is discovered, as well as a singe ensemble learning model to obtain the popularity scores, for social media prediction challenge.
- 3rd Workshop on Maritime Computer Vision (MaCVi) 2025: Challenge Results14
A comprehensive overview of the findings from the challenges addressed in the 3rd Workshop on Maritime Computer Vision 2025 is offered, evaluating trends from over 700 submissions.
- Impact of preoperative frailty on the surgical and survival outcomes in older patients with solid cancer after elective abdominal surgery14
Pretreatment frailty was significantly associated with worse survival outcomes and more surgical complications, with prolonged hospital stay, in the older patients with cancer after elective abdominal surgery in Taiwan.
- Predictive Value of Albumin and Neutrophil-to-Lymphocyte Ratio Score for Treatment Completeness and Safety Profiles in Patients With Head and Neck Cancer Receiving Definitive Concurrent Chemoradiotherapy13
The study showed that the ANS can accurately predict the treatment completeness of CCRT in patients with HNC and can be widely used as a simple predictor of treatment tolerance and safety profiles in patients without head and neck cancer.
- 13
- Multi-head self-attention mechanism enabled individualized hemoglobin prediction and treatment recommendation systems in anemia management for hemodialysis patients12
A neural networks based on multi-head self-attention mechanisms in an efficient and effective hemoglobin prediction model that holds great potential for individualized anemia management as a computerized clinical decision support system for hemodialysis patients.
- 12
- 12
- 11
- A Comprehensive Study of Spatiotemporal Feature Learning for Social Medial Popularity Prediction11
By reformulating the SMP tasks and integrating the multi-modal feature aggregation to the base learner for better performance, how identity-preserving is an essential property for SMP task is discussed and how temporal features' importance is evaluated.
- Adjuvant Chemoradiotherapy Associated with Improved Overall Survival in Resected Esophageal Squamous Cell Carcinoma after Neoadjuvant Chemoradiotherapy in Intensity-Modulated Radiotherapy Era11
It is demonstrated that adjuvant CRT was independently associated with a significantly improved survival and lower risk of recurrence than observation in esophageal squamous cell carcinoma patients staged as ypT3 and/or ypN+ after receiving neoadjuvant chemoradiotherapy and radical surgery.
- 11
- DenseSR: Image Shadow Removal as Dense Prediction10
This framework uniquely synergizes two key strategies: (1) deep scene understanding guided by geometric-semantic priors to resolve ambiguity and implicitly localize shadows, and (2) high-fidelity restoration via a novel Dense Fusion Block in the decoder.
- Spatiotemporal Feature Learning Based on Two-Step LSTM and Transformer for CT Scans10
A novel, effective, two-step-wise approach to tickle this issue for COVID-19 symptom classification thoroughly and it is suggested that the model ensemble could be adopted for more stable and promising performance in real-world applications.
- Dual Reconstruction with Densely Connected Residual Network for Single Image Super-Resolution10
A simple strategy of adding more shortcuts in the proposed network enables a faster learning process as the gradient information can be back-propagated more easily and achieves excellent performance in the real-world image super-resolution challenge.
- Integrating Frailty Assessment to Enhance Care in Cancer Patients with Borderline Eastern Cooperative Oncology Group Performance Status9
The findings highlight the multifaceted nature of patients with an ECOG PS of 2 and emphasize the importance of frailty assessment for treatment outcomes.
- Revisiting Vision-Language Features Adaptation and Inconsistency for Social Media Popularity Prediction9
This paper critically examines the applicability of CLIP-based features in SMP prediction, focusing on the overlooked phenomenon of semantic inconsistency between images and text in social media posts, and finds that incorporating inconsistency measures and adapted text features significantly improves model performance.
- Frailty assessment by two screening instruments in non-elderly patients with head and neck cancer9
Frailty assessment using fTRST and mFI-5 is valuable for predicting treatment-related adverse events, treatment tolerance, and QoL in non-elderly patients with HNC and only GA showed significant prognostic value for overall survival.
- 9
- 9
- 8
- HyFusion: Enhanced Reception Field Transformer for Hyperspectral Image Fusion8
HyFusion, a novel Dual-Coupled Network (DCN) framework designed to enhance cross-domain feature extraction and enable effective feature map reusing, provides a practical and effective solution for HSI fusion in resource-constrained scenarios, setting a new benchmark in hyperspectral imaging.
- 8
- COVID-19 Infection Percentage Estimation from Computed Tomography Scans: Results and Insights from the International Per-COVID-19 Challenge7
An overview of the COVID-19 infection percentage estimation challenge (Per-COVID-19) held at MIA-COVID-2022 is provided, to test the efficiency of modern deep learning methods on COVID-19 infection percentage estimation (CIPE) from CT scans.
- 7
- 7
- 7
- Comprehensive genomic profiling in multiple cancer types: A comparative analysis of the National Biobank Consortium of Taiwan and clinical practice cohorts6
This study provides comprehensive insights into the genomic profiles of diverse cancers in Taiwan, highlighting the crucial role of CGP in identifying actionable genomic alterations and guiding effective therapeutic strategies in real-world practice.
- Self-supervised Fusarium Head Blight Detection with Hyperspectral Image and Feature Mining6
This study proposes a self-unsupervised classification method based on HSI endmember extraction strategy and top-K bands selection, designed to analyze material signatures in HSIs to derive discriminative feature representations, making it more suitable for practical uses.
- 6
- QuantTune: Optimizing Model Quantization with Adaptive Outlier-Driven Fine Tuning5
QuantTune is proposed, a quantization-friendly fine-tuning method based on restricting the dynamic range amplification effect of outliers across Transformer-based models using the proposed outlier-driven loss, which achieves significant improvements across Transformer-based models, using naive PTQ only.
- CSAKD: Knowledge Distillation with Cross Self-Attention for Hyperspectral and Multispectral Image Fusion5
A novel knowledge distillation (KD) framework for HR-MSI/LR-HSI fusion to achieve SR of LR-HSI, and a novel Cross Self-Attention (CSA) fusion module to adaptively fuse those features to improve the spatial and spectral quality of the reconstructed HR-HSI.
- GRACE: Graph-Regularized Attentive Convolutional Entanglement with Laplacian Smoothing for Robust DeepFake Video Detection5
This paper introduces a novel method for robust DeepFake video detection, harnessing the power of the proposed Graph-Regularized Attentive Convolutional Entanglement (GRACE) based on the graph convolutional network with graph Laplacian to address the aforementioned challenges.
- 5
- 5
- Strong Baseline and Bag of Tricks for COVID-19 Detection of CT Scans5
A novel slice selection method is proposed for each CT dataset to address compatibility issues and effectively filtering out uncertain slices and enhancing the model's performance, ultimately achieving promising results.
- Potential Utility of the Mallampati Score for Prediction of Treatment Compliance and Safety Profiles of Patients With Head and Neck Cancer Undergoing Definitive Concurrent Chemoradiotherapy5
The results support the utility of Mallampati score as a surrogate for measuring RMO to predict survival outcomes, treatment compliance, and safety profiles in patients with HNC undergoing CCRT.
- Self-Learning-Based Low-Quality Single Image Super- Resolution5
A self-learning-based SR framework to simultaneously achieve low-quality single-image SR and compression artifact removal and deblocking can be simultaneously achieved via sparse representation and MCA (morphological component analysis)-based image decomposition.
- UMCL: Unimodal-generated Multimodal Contrastive Learning for Cross-compression-rate Deepfake Detection4
This work proposes a novel Unimodal-generated Multimodal Contrastive Learning (UMCL) framework for robust cross-compression-rate (CCR) deepfake detection, and achieves superior performance across various compression rates and manipulation types, establishing a new benchmark for robust deepfake detection.
- 4
- A Closer Look at Spatial-Slice Features Learning for COVID-19 Detection4
An enhanced Spatial-Slice Feature Learning (SSFL++) framework specifically designed for CT scan is introduced, enabling us to select crucial spatial slices for analysis by reducing 70% redundancy totally and proposed Kernel-Density-based slice Sampling (KDS) method to improve the stability during training and inference stage.
- Simple 2D Convolutional Neural Network-based Approach for COVID-19 Detection4
An advanced Spatial-Slice Feature Learning (SSFL++) framework specifically tailored for CT scans is proposed, which aims to filter out out-of-distribution data within the entire CT scan, allowing us to select essential spatial-slice features for analysis by reducing data redundancy by 70\%.
- 4
- 4
- 4
- 3
- A comprehensive study of age-related macular degeneration detection3
It is verified that different learning strategies need to be selected for the AMD symptoms segmentation tasks with varying characteristics of database, and demonstrated that current medical datasets suffer from annotation quality uncertainty, leading to limited learning capabilities.
- 3
- 3
- Multi Source COVID-19 Detection via Kernel-Density-based Slice Sampling2
The results demonstrate the effectiveness of the KDS-based pipeline on multi-source data and highlight the importance of dataset balance in multi-institutional medical imaging evaluation.
- Validation of a chemotherapy toxicity prediction model in older adults with cancer in Taiwan.2
This prognostic study validated the mCARG model in a heterogeneous cancer cohort in Taiwan but failed to do so in the lymphoma subgroup, suggesting mCARG could offer insights into patient survival among older individuals with cancer.
- 2
- Effect of prophylactic tube feeding in head and neck cancer patients with high Mallampati score undergoing definitive concurrent chemoradiotherapy2
Prophylactic tube feeding was associated with better treatment tolerance, safety profiles, and quality of life in patients with HNSCC and high Mallampati scores who underwent CCRT.
- Spatial-Slice Feature Learning Using Visual Transformer and Essential Slices Selection Module for COVID-19 Detection of CT Scans in the Wild2
An essential slices set algorithm is proposed to automatically select a subset of the CT scan, which could effectively remove the uncertain slices as well as improve the performance of the proposed SSFL.
- 2
- Adapting Object Detection to Fisheye Cameras: A Knowledge Distillation with Semi-Pseudo-Label Approach2
A lightweight object detection system custom-designed for fisheye cameras and optimized for quick deployment on embedded systems, combining knowledge distillation with semi-pseudo-label semi-supervised learning and a lightweight design perfect for real-time applications on constrained devices is introduced.
- COVID-19 Infection Percentage Prediction via Boosted Hierarchical Vision Transformer2
It is claimed that multi-tasking learning is beneficial for better feature representation learning for the infection score prediction and the maximum-rectangle cropping strategy is proposed to obtain the region of interest (ROI) to boost the effectiveness of the infection percentage estimation of COVID-19.
- 2
- 2
- 2
- 2
- Robust Hyperspectral Image Panshapring via Sparse Spatial-Spectral Representation1
A novel framework for hyperspectral image pansharpening that effectively combines low-resolution hyperspectral images (LRHSI) with high-resolution multispectral images (HRMSI) through sparse spatial-spectral representation through sparse spatial-spectral representation is introduced.
- Progressive Alignment with VLM-LLM Feature to Augment Defect Classification for the ASE Dataset1
This work proposes a novel approach leveraging vision-language models (VLM) and large language models (LLM) for their zero-shot capabilities for multimodal defect classification, and introduces the ASE dataset, which includes rich textual descriptions alongside images, enabling multimodal defect classification.
- Validation of WHO 2017 Classification and Identification of Prognostic Factors in Patients with Pancreatic Neuroendocrine Neoplasms: A Real-World Experience in Taiwan1
The findings validate the prognostic utility of the WHO 2017 classification, particularly in differentiating NET G3 from NEC, and provides valuable guidance for clinical decision-making and treatment selection in patients with PanNETs.
- 1
- Anchoring Trends: Mitigating Social Media Popularity Prediction Drift via Feature Clustering and Expansion1
An Anchored Multi-modal Clustering and Feature Generation framework that discovers temporally-invariant patterns across data distributions, and leverages Large Language Models to generate semantic Anchor Features that enable prediction based on stable principles rather than ephemeral signals.
- 1
- OCR is All you need: Importing Multi-Modality into Image-based Defect Detection System1
An external modality-guided data mining framework to extract statistical features from images as a second modality to enhance performance, termed OANet (Ocr-Aoi-Net), and introduces feature refinement and a gating function in the OANet to optimize the combination of these features, enhancing inference and decision-making capabilities.
- MISS: Memory-efficient Instance Segmentation Framework By Visual Inductive Priors Flow Propagation1
MISS leverages visual inductive prior flow propagation, integrating intrinsic prior knowledge from the Synergy-basketball dataset at various stages: data preprocessing, augmentation, training, and inference, demonstrating commendable performance in scenarios characterized by limited data availability and memory constraints.
- Augment Before Copy-Paste: Data and Memory Efficiency-Oriented Instance Segmentation Framework for Sport-scenes1
Memory effIciency inStance Segmentation framework based on visual inductive prior flow propagation that effectively incorporates inherent prior information from the dataset into both the data preprocessing and data augmentation stages, as well as the inference phase is proposed.
- 1
- 1
- 1
- 1
- 1
- 1
- 1
- 1
- 1
- S$^{3}$RNet: Sparse Spatial–Spectral Representation With Hybrid Knowledge Distillation for Efficient Multispectral and Hyperspectral Image Fusion–
The Sparse Spatial-Spectral Representation Network (S$3}$RNet delivers state-of-the-art fusion accuracy, strong robustness to various noise levels, and competitive runtime, making it well suited for practical remote sensing applications that require efficient MS–HS image fusion.
- Commit Locally, Exit Globally: Coordinating Adaptive Sampling and Early Exit in Diffusion Language Models–
A training-free, candidate-aware early-exit framework that keeps the two axes separate and matches each decision to evidence of its own scope is introduced, referred to as LATCH ( Localized Acceleration with Tracked-Candidate Halting ).
- –
- –
- VISTA: Validation-Guided Integration of Spatial and Temporal Foundation Models with Anatomical Decoding for Rare-Pathology VCE Event Detection -- after competition results–
This work proposes VISTA, a metric-aligned multi-backbone framework for the RAREVISION task, which combines EndoFM-LV for temporal context and DINOv3 ViTL/16 for frame-level visual semantics, followed by a Diverse Head Ensemble (DHE), Validation-Guided Weighted Fusion (VGWF), and Anatomy-Aware Temporal Event Decoding (ATED).
- VISTA: Validation-Guided Integration of Spatial and Temporal Foundation Models with Anatomical Decoding for Rare-Pathology VCE Event Detection–
This framework combines two complementary backbones, EndoFM-LV for local temporal context and DINOv3 ViT-L/16 for strong frame-level visual semantics, followed by a Diverse Head Ensemble, Validation-Guided Hierarchical Fusion, and Anatomy-Aware Temporal Event Decoding to produce stable event predictions.
- –
- –
- –
- –
- GTATrack: Winner Solution to SoccerTrack 2025 with Deep-EIoU and Global Tracklet Association–
GTATrack is presented, a hierarchical tracking framework that win first place in the SoccerTrack Challenge 2025 and significantly reduced false positives to 982, demonstrating state-of-the-art accuracy in fisheye-based soccer tracking.
- –
- Real-world insights from comprehensive genomic profiling across multiple cancer types.–
This study provides comprehensive insights into genomic profiling across diverse cancers in Taiwan, highlighting the crucial role of CGP in identifying actionable genomic alterations and guiding effective therapeutic strategies in real-world practice.
- –
- –
- –
- –
- Multimodality-aware Intra-dialytic Hypotension Classifier: A Bilingual NLP Approach to Classify Dialysis Records–
A Multi-modality-aware IDH Classifier (MMIDHC) using a natural language processing (NLP) model aimed at the classification of English-Chinese bilingual dialysis records concerning IDH is developed.
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
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 2026-10-15. 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.