Stray Animal Adoption Platform

Update time: 2025-06-29

A smart stray animal adoption platform that integrates LINE BOT and web technologies, combining image recognition, large language models (LLMs), and database + map visualization. Users can report stray animals, match potential adopters, and interact with a personalized rich menu via LINE.

Environment

• Python Version：3.12.9
• OS: Windows 10/11
• Package Source：requirements.txt

Setup

Create Venv

python -m venv venv

Activate the Environment

• CMD User：

venv\Scripts\activate

• PowerShell User：

.\venv\Scripts\Activate.ps1

Install Package

pip install -r requirements.txt

Configuration Before Use

Before running the application, open config.ini and complete the following fields:

[LINE]
CHANNEL_ACCESS_TOKEN = <your channel access token>
CHANNEL_SECRET = <your channel secret>

...

If you plan to use the LINE BOT feature with a personalized rich menu, run the following script first:

python register_richmenu.py

Run the App

python app.py

Features

📷 Image Recognition: Identify whether uploaded images are dogs, cats, or other.

🤖 LLM Integration: Automatically generate promotional descriptions using large language models.

📱 LINE BOT: Report and search for animals via personalized interaction and menu.

🗺️ Web Platform: Search shelters, view data dashboards, and visualize reports on a map.

🐶 Animal Matching: Discover and save potential animals for adoption.

🧩 Modular Design: Clean separation of backend logic, handlers, and utilities.

Dataset

This platform combines the Oxford-IIIT Pet Dataset with our self-collected dataset, serving as the foundation for training and testing, supporting stray animal detection and shelter data integration:

• Oxford-IIIT Pet Dataset

  • 37 breeds in total: 12 cat breeds + 25 dog breeds
  • Each image includes bounding boxes and segmentation masks
  • Suitable for classification and detection tasks

• Self-Collected Dataset

  • Contains images of mixed-breed cats and dogs (stray animals) for closer alignment with real-world adoption cases
  • Adds a new bird category to extend recognition beyond cats and dogs
  • All images are manually annotated with bounding boxes in YOLO format

• Usage in This Platform

  • The YOLO model is trained on the following classes:
    • Mixed-breed major categories: cat, dog, bird
    • Oxford-IIIT Pet Dataset fine-grained categories (37 breeds):
      Abyssinian, Bengal, Birman, Bombay, British_Shorthair, Egyptian_Mau, Maine_Coon, Persian, Ragdoll, Russian_Blue, Siamese, Sphynx, american_bulldog, american_pit_bull_terrier, basset_hound, beagle, boxer, chihuahua, english_cocker_spaniel, english_setter, german_shorthaired, great_pyrenees, havanese, japanese_chin, keeshond, leonberger, miniature_pinscher, newfoundland, pomeranian, pug, saint_bernard, samoyed, scottish_terrier, shiba_inu, staffordshire_bull_terrier, wheaten_terrier, yorkshire_terrier

📌 Official dataset source: Oxford-IIIT Pet Dataset

Directory Structure

.
├── app.py                      # Main Flask server entry
├── config.ini                  # Configuration file (LINE & DB)
├── register_richmenu.py       # Script to register LINE rich menu
├── requirements.txt           # Python dependency list
├── README.md                  # Project documentation
│
├── data/                      # (Optional) directory for external data
├── db/                        # SQLite database files
│
├── src/                       # Core backend logic
│   ├── db_utils.py            # DB utility functions
│   ├── init_db.py             # DB schema and init script
│   ├── linebot_app.py         # Main LINE BOT logic
│   ├── line_menu.py           # Rich menu generator
│   ├── login.py               # (Optional) login system
│   ├── saved_animals.py       # Saved animal handler
│   ├── settings.py            # Read config.ini
│   ├── User.py                # User structure
│   ├── handlers/              # Event handlers (e.g., postbacks)
│   └── utils/                 # Tools
│
├── static/                    # Static web resources (CSS, JS, images)
│   └── questions.json             # JSON data for UI or quizzes
└──  templates/                 # HTML templates (e.g., index.html)