Phasuwut
ExperienceProjectsSkillsPublicationsAI ChatContactHire Me
Phasuwut

Full Stack · AI Engineer · Thailand

Portfolio

  • Work Experience
  • Projects
  • Skills
  • Publications

More

  • Certification
  • Demo
  • AI Chat
  • Contact

Download

  • Profile
  • Resume
  • CV

© 2026 Phasuwut Chunnapiya

phasuwut.job@gmail.com

Real-Time Smart Home Surveillance System for Parcel Delivery and Fence-Climbing Intrusion Detection

Extended smart home monitoring with segmentation, tracking, and instant intrusion alerts.

 Home AI Resume Chat Work Experience All Projects Demo Technical Skills Certification Publications Contact
Personal ProjectsYear 2026

Project Overview

Objective

Extended smart home monitoring with segmentation, tracking, and instant intrusion alerts.

Stack

YOLOv9-segByteTrackTelegram Bot API

Delivery highlights

  • Developed a real-time smart home surveillance system using YOLOv9-seg for instance segmentation and ByteTrack for multi-object tracking to monitor parcel deliveries and detect fenceclimbing intrusions. The system tracks detected objects across video frames and analyzes their movement relative to predefined boundary lines (e.g., fence areas). When a tracked object crosses the boundary, the system automatically triggers a Telegram alert notification, enabling homeowners to receive immediate real-time warnings and respond quickly to potential security events
Back to Topic ProjectsBack to All Projects

System workflow

How real-time surveillance alerts work

Live frames are segmented and tracked continuously; boundary rules convert suspicious movement into an immediate Telegram alert.

Rendering workflow diagram…

Scroll horizontally to explore the full workflow on smaller screens.

Related Projects

3 items

Parcel and Person Tracking Experiment Assets

Personal ProjectsYear: 2026

Used parcel-and-person media assets to explore persistent identities across video frames.

Parcel and Person Instant Segmentation

Personal ProjectsYear: 2026

Prepared parcel-and-person segmentation data and experimental outputs for comparing model families.

Small-Object Focused Augmentation for Electric Vehicle Charger Socket Detection and Tracking

Personal ProjectsYear: 2026

Built upon the previous Semi-Supervised Video Object Segmentation for Electric Vehicle Charger Socket Tracking project by applying small-object-focused data augmentation to address missed detections when charger sockets appeared smaller or farther from the camera. This improved the model’s ability to detect small and hard-to-detect sockets while reducing missed detections in real-world scenarios.