Technical laboratories and institutions still rely on paper-based visitor registries — slow, error-prone, and impossible to analyse. This project replaces that process entirely with an intelligent robot that identifies visitors by face, greets them by name, checks their temperature via a handshake sensor, and logs the visit automatically.
Hello Humans! Welcome to RoboVerse: An IoT Based Interactive Robot
14th International Conference on Computer Communication & Informatics (ICCCI) · January 2024
↗ View Paper & Certificate
How It Works
End-to-end visitor pipeline
1
Image Capture & Enhancement
The robot captures a visitor's image using the Pi Camera and applies OpenCV
preprocessing — brightness normalisation, noise reduction — to improve recognition accuracy.
2
Face Detection & Encoding
Detected face regions are converted to 128-dimensional feature embeddings (encodings)
stored in JSON format — enabling fast comparison without storing raw images.
3
SVM Classification
A Support Vector Machine compares the new visitor's embedding against the stored
database. Known visitors are identified by name; unknown visitors are flagged as guests. Achieved a
15% accuracy improvement over baseline.
4
Greeting & Temperature Check
The robot greets the identified visitor by voice using pyttsx3 and extends a
servo-powered handshake. The handshake mechanism integrates a temperature sensor — if the reading is normal,
entry is permitted; if abnormal, the visitor is cautioned.
5
Attendance Logging
Each verified visit is timestamped and logged automatically, replacing the manual paper
registry entirely.
Technology Stack
What powers it
OpenCV
Image capture, enhancement, face detection. 25% faster processing vs baseline.
SVM Classifier
Face embedding comparison. 15% accuracy improvement over simpler classifiers.
pyttsx3
Text-to-speech for personalised voice greetings.
Raspberry Pi
Onboard compute. Pi Camera Module for visual input.
Servo Motor
Physical handshake mechanism with integrated temperature sensor.
JSON Storage
Lightweight face encoding database. No raw image storage.
Outcome
Smart visitor management — no paper, no manual checks.
This project demonstrated that a full visitor management system — identification, health screening, and logging — can be built with commodity hardware and open-source tools. The research was accepted at IEEE ICCCI 2024, validating the approach of combining SVM-based face recognition with physical IoT interaction.