Afzlqyziovisyxss3wm5

徐立恆   Computer vision engineer

Research include Computer Vision, Deep Learning, Machine Learning , and Artificial Intelligence.

For person recognition, behavior recognition, neural networks, image training, indoor positioning. Drone recognition, computer vision algorithms and other fields have more practical experience and implement a number of AI projects.

Taipei City, TW   Mail: [email protected] 

Education

National Taiwan Ocean University (NTOU)  , Computer Science M.S. 2017 ~ 2019

Experience

1,2015 - 2016 

Implementation of the Ministry of Science and Technology College Student Program.

2,November  2016

Participate in the IT Application Team Competition Finals.

3,November 2016

Participate in the Yahoo E-Commerce App Group Competition Finals.

4,September 2017 - June 2019

Implementation of the Ministry of Science and Technology Artificial Intelligence (AI) project plan "Smart Drones AI Platform Development - Smart Flight Control (1/4)" (108-2634F-009-013)

5,November 2018 

Participate in the National Smart Manufacturing Big Data Analysis Competition. 

Ocean University representative

6,September 2017 - June 2019

Data Structure T.A 

Data Structure Programming T.A 

Data Mining and Practice T.A 

MATLAB Programming Design T.A 

HTML Programming T.A

7,August 2019

Paper

「A REAL-TIME DEPTH ESTIMATION APPROACH FOR MULTIPLE PERSONS USING SINGLE CAMERA AND OPENPOSE」

Accepted and published at the CVGIP Computer Vision Graphics and Image Processing

SKILL


Tools

TensorFlow

Pytorch

Caffe

Keras

Yolo2

OpenCV


Neural Network 

CNN

RNN

LSTM

GAN

Faster RCNN

AutoEncoder


Language

C++ 

Java

Python


Project 1  

Projects 01 00@2x


A REAL-TIME DEPTH ESTIMATION APPROACH FOR MULTIPLE PERSONS USING SINGLE CAMERA AND OPENPOSE

This research was supported by the Ministry of Science and Technology under Grant Number 108-2634F-009-013, and was managed by the Pervasive Artificial Intelligence Research (PAIR) Labs, Taiwan.


Use a single lens to calculate the actual distance for multiple persons. The results are applied to "image occlusion", "aerial photography" and "multi-person distances real-time display". 

It has been applied to the fields of drones, indoor positioning, and behavior detection.

Project 2

Projects 01 00@2xDEEP LEARNING SYSTEM WITH SELF-REGULATION BEHAVIOR


Use a phased neural network to improve the training of existing objects. 

Using CNN to perform batch training and feature screening for objects, thereby improving the recognition of objects.

Powered by CakeResumePowered by CakeResume
Powered by CakeResumePowered by CakeResume