AI Driven Smart Shopping System

Smart Shopping
Objective

Implement a smart shopping system where face recognition, action recognition, and object recognition with gallery-based product matching all work together. The system must detect and log shopper interactions accurately by combining these three components.

Work to be Done
1. Face Recognition

Recognize and identify shoppers at entry and exit.

Provide enrollment functionality for new shoppers.

2. Action Recognition

Detect shopper actions:

Reach to shelf

Picking up item

Inspecting product

Placing back item

Idle / no action

3. Object Recognition with Gallery

Match detected products against a gallery of images.

Support uploading multiple photos (front, back, sides) per product.

Recognize products during shopping and return product ID.

4. Combined Workflow

When a shopper interacts with products:

Face recognition identifies who the shopper is.

Action recognition detects what action is happening.

Object recognition identifies which product is involved.

All three must run together in real time to produce a single combined event.

5. Event Logging

Generate logs in the format:

{ shopper_id, action, product_id, timestamp }

Ensure each event links the correct shopper, action, and product.

6. Dataset Collection

Freelancer is responsible for gathering and preparing datasets required for training all three modules.

Include gallery dataset preparation workflow.

Output Requirement

Working codebase with models and APIs.

Real-time integration of face, action, and object recognition.

Event log output as JSON or database entry.