陳映竹

Aged: 24

Gender: female

Birthday: 1997-05-26


Phone: 0916-209-822

Email: [email protected]

Github: https://github.com/Ying-Chu-Chen


Education

 

Master of Science in Public Health, National Defense Medical Center

Advanced fields: Deep Learning, Data visualization, Statistics, Bioinformatics

2019 - 2021

Bachelor of Science in Public Health, Kaohsiung Medical University

Advanced fields: Biostatistics, Epidemiology, Genetics

2015 - 2019

Project Experience

Feb 2021 - Jun 2021

Assistive Diagnosis of Posterior Circulation Ischemic Stroke

Object segmentation, 3D convolutionStatistics

This project aimed to develop an auxiliary-annotating efficient tool based on a deep learning model, to provide a novel quantitative scale and improve the accuracy of posterior circulation ischemic stroke patients' prognosis. In addition, it helps physicians shorten the time for evaluating patients' prognoses from 30 minutes to less than 5 minutes.

PPT

Jan 2021 - Apr 2021

Face detection and identity recognition

Object detection, Verification loss , Apps 

We detect faces per image by object detection model, and use verification loss on published journal to quantify the similarities among multiple faces for recognizing each face's identity. Also, we combine models in Shiny Apps, users can create identity data in the database and then our App can automatically detect users' faces in the other photos. 

PPTVideoAPP

Expertise


  • Object segmentation
  • Object detection
  • Image classification
  • Statistics

Software


  • R (MxNet / Shiny-Apps)

Appendix of projects

Face detection and identity recognition

The model can detect multiple faces in photos and recognize facial identities. Pictures above show that different identities are represented by different colors, and their labels also show above.



Users can create identity data for the face database, upload images and enter facial identities, and then our App can automatically detect users' faces in the other photos without training the model again.





Assistive Diagnosis of Posterior Circulation Ischemic Stroke


The picture above shows a part of the comparison between model prediction and ground truth, and the Intersection over Union (IoU) of nine brain structures are all higher than 0.7. 


The results of the DeLong test showed that the ROC for predicting prognosis using our novel quantitative integrated score was significantly better (p = 0.035) than existing semi-quantitatively calculated scores.







Autobiography

Personality:

Communication: during the master's degree, I went to the hospital to communicate with multiple doctors, understand the clinical needs and discuss the feasibility of model application for cooperation. During my tenure as the captain of the school team, I led 40 team members, understood the needs of each team member, discussed with the coach and adjusted continuously, and learned the importance of listening and team centripetal force. In addition, I coordinated the planning of the preparatory work and process of interscholastic athletics.

Autonomous learning, exploration, and thinking: in developing deep learning projects, continue to read scholarly journals and books, absorb the advantages of journals for tasks and apply them in the project so that better models can be constructed and continuously optimized. Such an experience made me learn to seek resources independently, spend time exploring and learning, and find the ability to solve problems.

Learning new expertise: for each task or when working with others, I am not limited to what I have learned in the past and have a high degree of cooperation. I am willing to devote myself to learning new programming languages or knowledge in the professional field, to continuously enrich myself.

School resume:

I studied in the Department of Public Health during university and became interested in statistical analysis and epidemiology. I started to learn the SAS programming language and analyzed large-scale data such as the gene database and the health care database. When I applied for the research institute, I discovered that the National Defense Medical School Public Health Research had planned programming language courses, so I decided to apply for the exam and enter the school in the first place.

During the research period, I was exposed to the field of artificial intelligence due to programming language courses and learned that the application of artificial intelligence can bring many benefits to the medical system and in life. Under the recommendation of the professor, take the deep learning course of the doctoral class to expand the basic cognition of image classification, object detection, natural language processing, and image generation. Through continuous learning, I completed the development of two deep learning projects and won the thesis award.

Summarize: although I did not graduate from an undergraduate department, I have the ability to learn autonomously. I am willing to devote my energy and time to various tasks and make myself grow. If I have this honor to join your company to serve, I will ask myself so that I can grow with time and experience and take responsibility for this job.

Powered by CakeResumePowered by CakeResume