Chih-Horng Wu

Dr. Chih-Horng Wu


Research Focus & Metrics

Academic Career Length

15Years (2012-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
11 March 2021 in Journal of Gastroenterology and Hepatology
Verified
Journal Article
Closed access
PDF file
TSTung‐Hung Su
Tung‐Hung Su
CWChih-Horng Wu
Chih-Horng Wu
JKJia-Horng Kao
Jia-Horng Kao

8 May 2022 in Computer Methods and Programs in Biomedicine
Verified
Journal Article
Open access
CHChiu-Han Hsiao
Chiu-Han Hsiao
PLPing-Cherng Lin
Ping-Cherng Lin
LCLi-An Chung
Li-An Chung
FLFrank Yeong-Sung Lin
Frank Yeong-Sung Lin
FYFeng-Jung Yang
Feng-Jung Yang
TSTzu-Lung Sun
Tzu-Lung Sun
This paper proposes an encoder-decoder architecture for kidney segmentation. A hyperparameter optimization process is implemented, including the development of a model architecture, selecting a windowing method and a loss function, and data augmentation. The model consists of EfficientNet-B5 as the encoder and a feature pyramid network as the decoder that yields the best performance with a Dice score of 0.969 on the 2019 Kidney and Kidney Tumor Segmentation Challenge dataset. The proposed model is tested with different voxel spacing, anatomical planes, and kidney and tumor volumes. Moreover, case studies are conducted to analyze segmentation outliers. Finally, five-fold cross-validation and the 3D-IRCAD-01 dataset are used to evaluate the developed model in terms of the following evaluation metrics: the Dice score, recall, precision, and the Intersection over Union score. A new development and application of artificial intelligence algorithms to solve image analysis and interpretation will be demonstrated in this paper. Overall, our experiment results show that the proposed kidney segmentation solutions in CT images can be significantly applied to clinical needs to assist surgeons in surgical planning. It enables the calculation of the total kidney volume for kidney function estimation in ADPKD and supports radiologists or doctors in disease diagnoses and disease progression.

1 February 2014 in Journal of Hepatology
Verified
Journal Article
Closed access
YSYu-Yun Shao
Yu-Yun Shao
CWChih-Horng Wu
Chih-Horng Wu
LLLi-Chun Lu
Li-Chun Lu
SCSoa-Yu Chan
Soa-Yu Chan
YMYu-Yi Ma
Yu-Yi Ma
LCLi-Tzong Chen
Li-Tzong Chen

20 March 2014 in Journal of Gastroenterology and Hepatology
Verified
Journal Article
Closed access
PDF file
CWChih-Horng Wu
Chih-Horng Wu
MHMing-Chih Ho
Ming-Chih Ho
YJYung-Ming Jeng
Yung-Ming Jeng
CHChao‐Yu Hsu
Chao‐Yu Hsu
PLPo‐Chin Liang
Po‐Chin Liang
TSTiffany Ting-Fang Shih
Tiffany Ting-Fang Shih

9 May 2022 in Computer Methods and Programs in Biomedicine
Verified
Journal Article
Open access
CHChiu-Han Hsiao
Chiu-Han Hsiao
TSTzu-Lung Sun
Tzu-Lung Sun
PLPing-Cherng Lin
Ping-Cherng Lin
TPTsung-Yu Peng
Tsung-Yu Peng
YCYu-Hsin Chen
Yu-Hsin Chen
YHYennun Huang
Yennun Huang
Previously, doctors interpreted computed tomography (CT) images based on their experience in diagnosing kidney diseases. However, with the rapid increase in CT images, such interpretations were required considerable time and effort, producing inconsistent results. Several novel neural network models were proposed to automatically identify kidney or tumor areas in CT images for solving this problem. In most of these models, only the neural network structure was modified to improve accuracy. However, data pre-processing was also a crucial step in improving the results. This study systematically discussed the necessary pre-processing methods before processing medical images in a neural network model. The experimental results were shown that the proposed pre-processing methods or models significantly improve the accuracy rate compared with the case without data pre-processing. Specifically, the dice score was improved from 0.9436 to 0.9648 for kidney segmentation and 0.7294 for all types of tumor detections. The performance was suitable for clinical applications with lower computational resources based on the proposed medical image processing methods and deep learning models. The cost efficiency and effectiveness were also achieved for automatic kidney volume calculation and tumor detection accurately.

1 October 2017 in Journal of the Formosan Medical Association
Verified
Journal Article
Open access
CWChiueng-Fang Wu
Chiueng-Fang Wu
YWYu-Jen Wu
Yu-Jen Wu
PLPo-Chin Liang
Po-Chin Liang
CWChih-Horng Wu
Chih-Horng Wu
SPShinn-Forng Peng
Shinn-Forng Peng
HCHung-Wen Chiu
Hung-Wen Chiu
Radiofrequency ablation (RFA) provides an effective treatment for patients who exhibit early hepatocellular carcinoma (HCC) stages or are waiting for liver transplantation. It is important to assess patients after RFA. The goal of this study was to build artificial neural network models with HCC-related variables to predict the 1-year and 2-year disease-free survival (DFS) of HCC patients receiving RFA treatments. This study was a retrospective study that tracked HCC patients who received computer tomography-guided percutaneous RFA between January 2009 and April 2012. The numbers of total patients with 1-year and 2-year DFS were 252 and 179, respectively. A total of 15 HCC clinical variables were collected for the construction of artificial neural network models for DFS prediction. Internal validation and validation conducted using simulated prospective data were performed. The results showed that the model with 15 inputs showed better performance compared with the models including only significant features. Parameters for performance assessment of 1-year DFS prediction were as follows: accuracy 85.0% (70.0%), sensitivity 75.0% (63.3%), specificity 87.5% (71.8%), and area under the curve 0.84 (0.77) for internal validation (simulated prospective validation). For 2-year DFS prediction, the values of accuracy, sensitivity, specificity, and area under the curve were 67.9% (63.9%), 50.0% (56.3%), 85.7% (70.0%), and 0.75 (0.72), respectively, for internal validation (simulated prospective validation). This study revealed that the proposed artificial neural network models constructed with 15 clinical HCC relevant features could achieve an acceptable prediction performance for DFS. Such models can support clinical physicians to deal with clinical decision-making processes on the prognosis of HCC patients receiving RFA treatments.