Boopro Technology logo
eKonobar DOO

eKonobar Vision: AI-Powered Staff Call Detection in Restaurants

eKonobar DOO

Product Company
Company Type
Split, Croatia
Location
Project work
Engagement Model
1 - 5 people
Team Size
Less than 3 Months
Duration
Less than $20K
Budget

About

eKonobar Vision is a proof-of-concept project focused on developing and validating a custom AI solution for detecting staff calls in restaurants using computer vision techniques to improve service efficiency.

Challenge, approach, and impact

Real-time Performance

The system needed to process video streams with minimal latency and high throughput to function effectively in a live restaurant setting.

Dynamic Environments

Restaurant settings vary widely in lighting (indoor/outdoor), movement, and the number of people visible in the camera frame.

Accuracy in Cluttered Scenes

Detecting the specific 'waiter call' hand gesture among other unrelated movements was critical.

Model Efficiency

The AI models had to be both accurate and lightweight enough to run efficiently on cloud infrastructure like AWS EC2 without requiring costly specialized hardware.

Evaluate Computer Vision Models

Benchmark various pre-trained hand gesture recognition models in realistic restaurant scenarios.

Develop PoC Application

Build a command-line tool capable of processing live video streams from security or IP cameras.

Implement Detection Logic

Integrate TensorFlow Lite Runtime and MediaPipe models for recognizing the specific waiter call gesture.

API Integration

Automatically trigger a specified API endpoint when a valid waiter call gesture is detected.

Configurable Parameters

Enable camera source, API endpoint, and other system settings to be adjusted via a YAML configuration file.

Real-world Testing

Validate the system using real IP cameras in both indoor and outdoor conditions, with multiple users and simultaneous hand-raise scenarios.

How we built

AI & ML Solutions
VoIP
IoT & Smart Devices
API
Cloud Computing
Integrations

Testimonials

Uroš S., Software Developer @ Boopro Technology

Boopro Technology

Verified Testimonial

eKonobar Vision was a great challenge - clear goal and real-world conditions. We managed to detect waiter calls in real time without relying on expensive hardware. Testing across different lighting setups and with multiple people in frame made it a true proof of concept.

Team structure

Client team

Ivan P's avatar

Ivan P

Product Owner

Daily point of contact

The client stakeholders at eKonobar DOO were working closely with the team at Boopro Technology

Agency team

2 x Core Development Team's avatar

2 x Core Development Team

Production

Similar Case Studies

Slide 1 of 0
Boopro Technology's logo

Chat with agency

Let's turn your ideas into reality—connect with our team.