曾永源
Yung-Yuan Tseng

Mail: [email protected]

Mob: 0958-189-993


Personality: Friendly, Positive, and Compassionate. Meanwhile, I'm the one who like to challenge things that I've never tried before.

Experience: 德律科技( Test Research Innovation, TRI) as Software Engineer

  • Improved existing  object detection model  for more precise bounding box.
  • Developed and researched defect detection models for Automated Optical Inspection (AOI) on PCB board.
  • Maintained existing program in C++ language for object detection AI model post-processing.
  • Transferred AI models in Python to C++ and implemented the algorithms in C++.


SKILLS

Programming Language


  • C/C++
  • Python
  • JavaScript
  • Swift

Deep Learning Knowledge


  • Keras, Tensorflow, Pytorch
  • OpenCV
  • Saliency Detection
  • Objection Detection/Segmentation

Language


  • Chinese
  • English (TOEIC: 690, 2020/5)

Education

2017 - 2020

國立台灣科技大學 (National Taiwan University of Science and Technique)

資訊工程系 (Computer Science and Information Engineering)

2013 - 2017

國立嘉義大學 (National Chiayi University)

資訊工程系 (Computer Science and Information Engineering)

Experience


Master Thesis

My master thesis was A Multi-Patch Aggregated Aesthetic Rating System Based on Eye Fixation. The reason I chose this topic was due to the interest in photography. I used deep learning to extract the abstract aesthetic features in images. Through analysis these abstract features, machine might learn the subjectivity feelings like human. According to my MS research, I'm familiar with the topic of deep learning, which includes:

  1. Object Detection
  2.  Object Segmentation
  3.  Human Like Visual Attention(Saliency Detection)
  4.  Aesthetic Assessment

My thesis was published on IEEE MACHINE LEARNING FOR SIGNAL PROCESSING (MLSP) conference.


Contest Experience

Contest Title: National Smart Manufacturing Big Data Analysis Competition held by Tunghai University.


Contest Result: Second Round (total is two rounds)


Model architecture: LSTM


Detail: We used Zero_Mean and analyzed the frequency of peak in the data as our pre-process method. 


Internship Experience

Compony Type: network communication equipment 

Period: 6 months.

Content: built network routing protocol RIP and OSPF for embedded equipment then created a visualize operation (Web) and command line control interface.

Project


Project Title: MOST Industry 4.0 project

Purpose: Update machines without renewed machines.

Content: Built AR App on mobile device.

Period: 1 year.

Details:

  1. Showing the information of processing components in AR. 
  2. Measure the length of lathe cutters with AR.
  3. Digitize machine processing log information. 
  4. Visualize virtual lathe cutters models.

Special Reward


Graduate from NCU: Award for Moral Education Achievement.

  

Powered by CakeResumePowered by CakeResume