Welcome to my page, I AM

Jihwan Lee

My interests are in the areas of

About

About Me

문제 정의에 필요한 설득력, AI를 활용한 문제 해결 능력을 기르기 위해 공부하고 있습니다.

  • Name: 이지환
  • Research Interests: Computer Vision,
    Generative Models
  • Address: Seoul, Korea
  • Email: wlghks7790@gmail.com

github icon  Github

Education

2018.03-2025.08 (expected)

Hanyang University

Industrial Engineering

BACHELOR OF SCIENCE

Awards

2024.09.06

제 4회 KOPIS 빅데이터 공모전 <서비스 개발> 부문

돋움상 (3등)

공연 포스터의 저작권 침해 예방과 창의적 창작을 위한 이미지 검색 및 생성 웹 서비스

수여 기관

(재)예술경영지원센터

2024.06.28

대한전시디자인학회 제36회 학술대회

우수논문상

포스터 디자인의 저작권 침해 탐지를 위한 고차원 벡터 데이터베이스 구현

수여 기관

대한전시디자인학회

2023.12.29

GDSC X 알파코 K-디지털 플랫폼 나무톤 2023

최우수상 (2등)

취업 준비생들을 위한 취업 준비 솔루션 웹 서비스

수여 기관

(주)알파코

2023.12.21

고려대학교 지능정보SW아카데미 3기 최종 프로젝트 평가

우수상 (3등)

저작권 침해 예방 및 인식 개선을 위한 저작권 이미지 유사도 분석 웹 서비스

수여 기관

한국데이터산업협회

2023.09.24

2023 제주 스마트관광 빅데이터 해커톤

최우수상 (2등)

시니어를 위한 수요 예측 기반 제주 관광지 추천 서비스

수여 기관

제주관광공사

Projects

2024.11.19~2024.12.12

Fine-Tuning (w/ LoRA) Stable Video Diffusion for Moon Phases Video

SVD Project Image AI Theory and Programming lecture
Enhanced the Stable Video Diffusion model by fine-tuning it with LoRA to generate videos illustrating the phases of the moon. Gained a deep understanding of the Diffusion concept and the principles of image generation, leveraging this knowledge to fine-tune generative AI models for both image and video creation.

 Tech stack

PyTorch, OpenCV, Diffusers, Pillow, Accelerate, Transformers

Main role

Team Leader, Fine-Tuning Code Implementation, Final Presentation (Fine-Tuning Section)

2024.09.02-2024.12.09

Ensemble of Sparsely connected RNN and Autoencoder for Anomaly Detection in Time Series

IE Project Image Industrial Engineering Capstone
Implemented Ensemble of Sparsely Connected RNN + AutoEncoder (AE) models for time series anomaly detection, based on insights from existing research. Customized modules for experimental purposes and incorporated advanced techniques such as Residual Connections and Attention mechanisms to evaluate and compare the performance of each enhanced model.

 Tech stack

PyTorch, Scikit-learn, Matplotlib, Numpy, Pandas, Seaborn

Main role

Team Leader, Model Implementation, Final Presentation

2024.08.29-2024.10.04

My Study Guide: Chrome Extension to maintain focus during their learning & browsing sessions

MSG Project Image Gemma Sprint (Google Machine Learning BootCamp 2024)
Implemented a chrome extension for user learning concentration analysis. My Study Guide analyzes your web browsing history based on the learning topics you set. It evaluates how closely each visited website relates to your specified topics, assigns a relevance score to each browsing record, and visualizes the results in a comprehensive report.

 Tech stack

Python, JS

Main role

Front-End (Chrome Extension) Implementation

2024.07.11-2024.09.06

POST-CO: Image Search & Generation Web Service for "Poster-Copyright"

Post-Co Project Image The 4th KOPIS Big Data Contest <Service Development> Section
This project was conceptualized based on retrospectives from the CHECK-IN project. Post-Co is a web service designed to prevent copyright infringement in performance posters while fostering creative content creation. This service allows users to upload images or text to search for related posters and provides tools to generate or modify images based on those results. Post-Co is tailored to help creators avoid copyright issues while producing original and high-quality content.

 Tech stack

PyTorch, OpenCV, Flask, MySQL, OpenAI API (DALL-E 3), FAISS, HTML, CSS, JS, Elastic Beanstalk, Naver Cloud Platform, NAVER PAPAGO API

Main role

Team Leader, Idea Planning, Client and Application Development, Web site Deployment, Final Presentation

2023.11.29-2024.02.22

Taek-it-Easy: Android App for people with intellectual disabilities to learn Taekwondo

Taek-It-Easy Project Image Google Solution Challenge 2024
Created education application, Taek-it-Easy which was in collaboration with Taekwondo experts who directly teach individuals with intellectual disabilities. Core features include learning postures through videos recorded by Taekwondo experts, receiving feedback by comparing user-recorded postures to model postures, and incorporating engaging elements such as music and praise stickers to encourage continuous learning.

 Tech stack

TensorFlow, MySQL, Pandas, OpenCV

Main role

Pose Estimation Model Research, Domain Knowledge Acquisition, DB Development, Video Data Preprocessing

2023.12.28-2023.12.29

WorkReadyFlow: Job preparation solutions web service for job seekers

WorkReadyFlow Project Image NAMUTHON: 2023 GDSC X Alpaco Corp. Hackathon
Designed an AI-powered service tailored for job seekers, offering the following features based on user-provided self-introduction essays: spell checking, word count analysis, and AI-generated interview questions. Additionally, planned a simulated video interview module utilizing AI-generated and company-specific interview questions to analyze responses, eye contact, and pronunciation for performance feedback.

 Tech stack

OpenAI API (davinci-002), Google Web Speech API, T5, Flask

Main role

Idea Planning, ML Model and API Research, API Server Implementation, Final Presentation

2023.09.01-2023.12.21

CHECK-IN: CHECK copyright INfringement

CHECK-IN Project Image Korea University INISW Academy 3rd Cohort, Final Project
Created CHECK-IN which is to prevent unintentional copyright infringement and raise awareness about copyright issues. Various indicators are used for this analysis, including cosine similarity, structural similarity index (SSIM), and text domain similarity.

 Tech stack

PyTorch, Scikit-learn, SQLite, Flask, Numpy, Pillow, OpenCV, HTML, CSS, JS

Main role

Team Leader, Idea Planning, Model Research, Web Service Implementation, Mid-term and Final Presentation

2023.09.22-2023.09.24

Jeju Tourist Spot Recommendation Service Based on Demand Forecast for Seniors

Jeju Project Image Jeju Smart Tourism Big Data Hackathon
Developed a demand prediction-based Jeju tourism recommendation service tailored for senior tourists. Conducted clustering of senior tourists and generated recommendation rules within each cluster by applying the Apriori algorithm to tourism category relationships. Additionally, utilized a Bi-LSTM model for demand forecasting on tourism entry data, enhancing the system to prioritize recommendations for attractions with predicted demand below the average within the recommended categories.

 Tech stack

PyTorch, Streamlit, Pandas, Numpy

Main role

Idea Planning, Modeling and Web Prototype Implementation, Final Presentation

Activities

2024.09. ~ 2024.12.

(Mentoring) Alpaco Corp. Youth Digital Problem Solving Project

Seoul Dongyang High School

AI bias analysis and visualization project using text data (24.09. ~ 24.11.)
Online Career Mentoring (24.11. ~ 24.12.)

2024.07. ~ 2024.10.

Google Machine Learning BootCamp 2024

Cohort 5th member

Coursera - Deep Learning Specialization (24.07. ~ 24.08.)
Gemma Sprint (24.08. ~ 24.10.): Implementing Chrome Extensions for Learning Concentration Analysis with Gemma 2

2023.09. ~ 2024.08.

GDSC - Hanyang

General Member

23-2 study: Learn AI thoery, Perform mini project about tracking football players
24-1 study: Review linear algebra, Research on AI Papers
side project: lms Lecture Summary + Summary Management Web Service Implementation, ML Part

2023.09. ~ 2023.12.

Korea University INISW Academy

Cohort 3rd Member

1st, study: DB, Web, Python, ML/DL theory and programming practice
2nd, project: Implement web service called CHECK-IN (CHECK copyright INfringement)

Contact

Contact Me

Address

Seoul, Korea

Email Address

wlghks7790@gmail.com

github icon

GitHub

abcd-EGH