Livoa LogoLivoa
Requirements
Data
Constraints
Optimiser parameters
Stop criteria
Metrics

Modelling error

Structural complexity

Robustness

Available models

Data preprocessing models

Machine learning models

User input

Feature selection
Data preprocessing
Models selection
Models tuning

Automated

design

GOLEM graph optimiser

Initial pipeline

Composite pipeline

Modelling results

Metrics evaluation

Pipeline serialization

Composing

User
Mobile
Desktop
Tablet
Frontend: React + Tailwind

Web Speech TTS · Accessibility features

Camera (camera capture multi-frame)
Voice
Interface (capture, results)
Local Storage
API ENDPOINTS

• POST /api/detect
• GET /api/products

Backend: Python Flask
Flask Server (Routing and endpoints)
Image Processing (resize / enhance / crop)
Product DB (products.json / Postgres)
Error Handling (Logging / fallback)
AI Engine

YOLOv8 (detection) + pyzbar (barcode)
Inputs: image frame | Outputs: product key, confidence.

Data Storage
products.json / PostgreSQL
yolov8x_best.pt (weights)
config and runtime data
user data / logs
Tech Stack

• Frontend: React, Tailwind
• Backend: Flask, ultralytics
• AI: YOLOv8, pyzbar, pytesseract
• DB: products.json / Postgres
• Dev: VS Code, Docker (optional)

Legend

Barcode = deterministic (high confidence)
YOLO = probabilistic (confidence score)
OCR = fallback (text match)

Copy of Project

by aizen

0
0 uses