Interests 研究興趣

研究重點領域

• Image processing 影像處理
• Machine learning 機器學習
• Computer vision 電腦視覺
• Deep learning 深度學習

研究方向

• A short introduction to my research: [PPT] [PDF] (Latest updated: Feb. 2018)
• 研究簡介：[PPT] [PDF]

COVID-19 Symptoms Detection in CT Scan

[IEEE ECCV Workshop 2022] [1st place in COV19D challenge!!]

Spatial-Slice Feature Learning using Visual Transformer and Essential Slices Selection Module for COVID-19 Detection of CT Scans in the Wild

[IEEE ICCV Workshop 2021] [3rd place in COV19D challenge!!]

Adaptive Distribution Learning with Statistical Hypothesis Testing for COVID-19 CT Scan Classification

Detecting the COVID-19 Symptoms of CT scan in the wild. Our model is robust to noise, as well as adapting to various spatial and slice resolutions.

Social Media Prediction as Longitudinal Task (2022-)

[ACM Multimedia 2022]

C.C. Hsu, P.J. Tsai, T.C. Yeh, and X.U. Hou, “A Comprehensive Study of Spatiotemporal Feature Learning for Social Medial Popularity Prediction,” 30th ACM international conference on Multimedia, 10-14 Oct. 2022.  [PDF]  

We form the SMP task as an identity-preserving time-series task to observe the popularity scores for the specific identity. In addition, more valuable and essential features could be explored.
In this paper, 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 tasks is discussed.

Semantic Segmentation for Autonomous Driving (2021-)

[IEEE ICME Workshop 2022]

Augmented-Training-Aware Bisenet for Real-Time Semantic Segmentation  [PDF]  

[IEEE ICASSP 2022]

DCSN: Deformable Convolutional Semantic Segmentation Neural Network for Non-Rigid Scenes  [PDF]  

As the requirements for autonomous driving rapidly grow, the critical issues here could be stable and fast inference so that driving could be more reliable and computationally cheap. We have proposed a novel deformable SS for more stable results for outdoor scenes. While the lightweight hardware becomes more attractive, we also proposed a fast segmentation method as well as keeping the comparable performance.

Fake Image/Video (DeepFake) Detection (2018-)

[IEEE ICIP Conference 2019] [Applied Sciences]

Detecting Generated Image Based on Coupled Network with Two-Step Pairwise Learning

[IEEE IS3C Conference 2018]

Learning to Detect Fake Face Images in the Wild

[News] 工商時報、台大新興媒體中心

[Project] [PDF]  [GitHub] [Also check our online demo]

偽造 / 造假 照片偵測!! 打擊假照片、假新聞！

Deep Compressed Sensing for Hyperspectral Images (2020-)

[IEEE Transactions on Geoscience and Remote Sensing]

DCSN: Deep compressed sensing network for efficient hyperspectral data transmission of miniaturized satellite [PDF]

 [CVGIP 2020]

Deep joint compression and super-resolution low-rank network for fast hyperspectral data transmission,

[Project] [GitHub] 

以深度學習為基礎之高光譜/多光譜影像超解析度與壓縮感知技術開發

Decision-making of Autonomous Vehicles using Vision Information (2019-)

 [Journal Multimedia Tools and Applications]

Deep Learning-based Vehicle Trajectory Prediction based on Generative Adversarial Network for Autonomous Driving Applications

[IEEE ICCE-TW 2020]

Learning to Predict Risky Driving Behaviors for Autonomous Driving

[Large-Scale Vehicle Collision Dataset @ TW] [Link]

自駕車視覺系統之危險駕駛行為預測與台灣道路地區資料庫建置。

Social Media Prediction (2016-)

• [ACM Multimedia 2017-2020]
• Social Media Prediction Based on Residual Learning and Random Forest (2017)  
  • See the publication list to check new versions.
• 2 * Best-performance award, 2 * Top-performance award
• Best Grand Challenge Paper Award (2017)
• [GitHub][PDF]

預測你的PO文點擊率!!

Identity-Preserving Face Hallucination (2018-2020)

[IEEE T-IP]

SiGAN: Siamese Generative Adversarial Network for Identity-Preserving Face Hallucination," IEEE Transactions on Image Processing (T-IP), 2019.

[PDF]  [GitHub]

還原不清楚、模糊的低解析度人臉照片，並保留原始身分！

Large-Scale Image Clustering (2016-2017)

[TMM 2018] [ICIP 2017]

[CNN-Based Joint Clustering and Representation Learning with Feature Drift Compensation for Large-Scale Image Data]

[PDF] [Code]

巨量影像資料分群演算法。

Image Deblocking and Super-Resolution (2013-2014)

[TMM 2015] [MMSP 2013]

[Learning-Based Joint Super-Resolution and Deblocking for a Highly Compressed Image]

*MMSP 2013 Top 10% paper award

[Project Page] [PDF] [Matlab Source Code (32-bit only)]

同時去除區塊效應並提高解析度 (放大後不模糊)!

Super-Resolution of Textured Video (2012~2014)

[IEEE Transaction on Image Processing (TIP)] [MMSP 2014]

[Temporally coherent super-resolution of textured video via dynamic texture synthesis]

[Project Page] [PDF] [Matlab Code]

提供動態紋理視訊的超解析度技術 (放大後不模糊)!

Quality Assessment for Image Retargeting (2011~2013)

[IEEE Journal of Selected Topics in Signal Processing] [VCIP 2013]

[Objective quality assessment for image retargeting based on perceptual geometric distortion and information loss]

[Project Page] [PDF] [Matlab Code]

評估影像濃縮技術的品質 (即對這張影像結果作評分)!

Super-Resolution (2010-2011)

[MMSP2011]

[Image super-resolution via feature-based affine transform]

[Project Page] [PDF] [Executable Code (Matlab)]

We provided only implements NLM with the proposed method as an example.

影像超解析度技術依賴於資料庫，我們提出一種方法豐富資料庫的類型，提高放大的效果。

Face Hallucination (2008-2010)

[MMSP2010]

[Face hallucination using Bayesian global estimation and local basis selection]

[Project page] [PDF] [Matlab code & Database]

人臉超解析度放大 (從很小張人臉影像放大到清晰的人臉影像)!

Video Forensics (2007-2008)

[MMSP2008]

[Video forgery detection using the correlation of noise residue]

*Citations > 100

[PDF] [Matlab code] [Database]

視訊的鑑識技術，防止有心人士對影片偽造。

Image Authentication (2006-2007)

[Optical Engineering]

[Image authentication and tampering localization based on watermark embedding in the wavelet domain]

[Download PDF] [Source Code]

將浮水印藏入影像中並可耐受不同攻擊。