Computer Vision 2025 Term Project Exhibition

Classification

• 반려동물 안구 질환 분류 프로젝트 [Github]
  • Author: 백종언
  • Method: ViT, ResNet50, Random Forest, Stacking Ensemble, XGBoost, *Web Deployment
• Food Nutrition Predictor [Github]
  • Author: 이승헌
  • Method: ResNet50, Food-101 Dataset (101 classes), *Web Deployment (Streamlit)
• 실시간 수어 해석 [Github]
  • Author: 양도희
  • Method: MediaPipe, LSTM
• Pneumonia Detector [Github]
  • Author: 차승우
  • Method: U-Net (masking), Random Forest
• Emotion Vision: Real-time Facial Emotion Recognition [Github]
  • Author: 마이 쥬이 타잉
  • Method: Custom CNN Network, FER2013 dataset
• My Digit Recognizer [Github]
  • Author: 홍대연
  • Method: Custom CNN
• Enhanced Facial Emotion Recognition System [Github]
  • Author: 신 폰
  • Method: Manual Feature Weighting

Segmentation and Detection

• Food Calories Estimation [Github] 🏆
  • Author: 박서현
  • Method: Mask2Former (fine-tunned with FoodSeg103 dataset), Gemini, *Web Deployment (gradio)
• FIND LOGO [Github]
  • Author: 최민기
  • Method: YOLOv8 (classes: logos)
• Smart Waste Sorter [Github]
  • Author: 테무렌
  • Method: YOLOv5 (classes: recyclable, non-recyclable), *Web deployment
• 도로 재난 실시간 탐지 프로젝트 [Github]
  • Authors: 강현, 곽용진
  • Method: YOLOv8 (classes: flood, fire, car accident, sink hole), Optuna (hyperparameter optimization)
• Recycling Classification Detector [Github]
  • Author: 박솔지
  • Method: YOLOv8 (classes: can, paper, plastic), MediaPipe
• Parking Management using YOLO [Github]
  • Author: 천쥐닷
  • Method: YOLOv11
• Mosaic Recorder [Github]
  • Author: 최민혁
  • Method: YOLOv8
• Testing SAM2 [Github]
  • Author: 손장우
  • Method: SAM2
• Person Detector [Github]
  • Author: 김종운
  • Method: YOLOv8
• 실시간 객체 인식 [Github]
  • Author: 신대혁
  • Method: MobileNet-SSD

Events and Actions

• Bento PT 1.0 [Github]
  • Author: 김태현
  • Method: YOLOv8x (human detection), YOLOv8x-pose (human pose estimation), ST-GCN (@MMAction2)
• Real-Time Fall Detection System [Github]
  • Author: 판 떤 다이
  • Method: YOLO, MediaPipe, *Telegram Alert, *Web Deployment (Streamlit)
• Real-time Lane Intrusion Detection and License Plate Recognition System [Github]
  • Author: 박지환
  • Method: YOLO, EasyOCR, EDSR (super resolution), CSRT Tracker (@OpenCV)
• AutoFine AI (불법 주정차 감지 및 AI 기반 자동 신고 시스템) [Github]
  • Author: 사라
  • Method: YOLOv8, ByteTrack
• Crosswalk Detection and Traffic Signal Assistance System [Github]
  • Author: 이스칸다로브 후산
  • Method: YOLOv8

AR and 3D

• AR Puzzle Game [Github]
  • Author: 신연식
  • Method: ArUco Tags
• 3D Depth Effect [Github]
  • Author: 정수민
  • Method: MiDaS (https://github.com/isl-org/MiDaS)
• MotionCapture [Github]
  • Author: 신해성
  • Method: MediaPipe, *PyOpenGL (for rendering)
• Dancing on the Chessboard [Github]
  • Author: 김동욱
  • Method: PnP, Homography
• 2D to AR [Github]
  • Author: 윤성현
  • Method: MediaPipe
• Person-Hall SfM & 3D Visualization [Github]
  • Author: 정민욱
  • Method: COLMAP, Open3D

Faces

• NALDA Timer [Github]
  • Author: 박성재
  • Method: Human Library (https://github.com/vladmandic/human), *Web Deployment
• 졸음 감지 프로그램 (Drowsiness Detection) [Github]
  • Author: 김동훈
  • Method: Face Detector (@OpenCV), Face Landmark Extractor (@dlib), Custom CNN Network (eye open/close classifier), PnP
• StudyFocus Project [Github]
  • Author: 전소원
  • Method: MediaPipe, *Web Deployment
• 눈 집중도 탐지기 [Github]
  • Author: 김현서
  • Method: Face Detector (@dlib), Face Landmark Detector (@dlib), *PyQt
• Driver Drowsiness Prevention System [Github]
  • Author: 이시헌
  • Method: MediaPipe
• sensingbox [Github]
  • Author: 박민서
  • Method: MediaPipe
• FaceTrack360: Multi-Face Pose and State Detection System [Github]
  • Author: 린 렛 윈 데인
  • Method: MediaPipe
• Blur Person Faces: 실시간 영상 인물 얼굴 모자이크 프로젝트 [Github]
  • Author: 이현빈
  • Method: RetinaNet (human detector), SSD (face detector)
• Eye State Detection [Github]
  • Author: 암갈란투르
  • Method: MediaPipe, Custom CNN Network

Hands and Gestures

• Finger Count Classifier [Github]
  • Author: 김나연
  • Method: MediaPipe, KNN Classifier (@scikit-learn)
• Silent Heart [Github]
  • Author: 신용준
  • Method: MediaPipe, Custom Neural Network (@scikit-learn)
• Hand Tracking 3D Drawer [Github]
  • Author: 강형석
  • Method: MediaPipe, PnP
• Sign Language Recognition [Github]
  • Author: 김강현
  • Method: Custom CNN Network, Sign Language MNIST dataset
• Hand-Controlled PC Interaction [Github]
  • Author: 김태견
  • Method: MediaPipe
• Gesture Image Viewer [Github]
  • Author: 이준수
  • Method: MediaPipe
• HandSketch [Github]
  • Author: 서준수
  • Method: MediaPipe
• AR [Github]
  • Author: 한결
  • Method: MediaPipe
• Gesture and Voice Controlled Mouse [Github]
  • Author: 표먓되
  • Method: MediaPipe

Games

• G-Beats (Gesture-Beats) [Github]
  • Author: 조강우
  • Method: MediaPipe, *Unity
• I Tracker (방탈출 게임) [Github]
  • Authors: 김민석, 김세원, 박재홍
  • Method: Face Detector (OpenCV), Face Landmark Extractor (dlib), *Unity
• ParkingYOLO [Github]
  • Author: 박종욱
  • Method: YOLOv8, *Web deployment
• ROCK-PAPER-MEOW [Github]
  • Author: 정수정
  • Method: MediaPipe, MobileNetV2 (classes: rock, paper, scissors)

Image Generation

• 스타일 변환기 [Github]
  • Author: 홍석현
  • Method: StyleTransfer
• Neural Style Transfer [Github]
  • Author: 나잉
  • Method: VGG19 backbone
• 자연스러운 이미지 생성 [Github]
  • Author: 자워키르베크
  • Method: StyleGAN3

Image Processing

• PythoEditor [Github]
  • Author: 김성현
  • Method: PyGame
• Red-Green-Assistant-AR [Github]
  • Author: 문석준
  • Method: RGB Thresholding
• Perspective Correction Tool [Github]
  • Author: 이민준
  • Method: Homography

Characters and Texts

• 수식 인식 칠판 OCR 프로젝트 [Github]
  • Authors: 심종우, 김지민
  • Method: ResNet18-based Classifier, *Notion Deployment
• Video Subtitle Translator [Github]
  • Author: 이동은
  • Method: EasyOCR, py-hanspell, MarianMT, TTS (https://pypi.org/project/TTS/)
• OCR Translator [Github]
  • Author: 이경민
  • Method: Tesseract OCR, deep-translator

Misc

• ASCII Video Player (터미널 기반 비디오 재생기) [Github]
  • Author: 정세영
  • Method: Intensity
• CNN 기반 자율주행 프로젝트 [Github]
  • Author: 이경빈
  • Method: A2C, *AirSim (https://microsoft.github.io/AirSim/)
• Hand Matching System (손 사진 비교로 동일 인물 판별하기) [Github]
  • Author: 이먓판
  • Method: SuperPoint, SuperGlue
• Smart Healthcare System [Github]
  • Author: 조현규
  • Method: MLP Classifier (@scikit-learn)
• R-Guardian (Ransomware Detector) [Github]
  • Author: 김성우
  • Method: Entropy
• Scanner [Github]
  • Author: 이채현
  • Method: Edge/Contour Extraction