PROJECTS
AIKO
A Social Companion Robot
A companion robot powered by Raspberry Pi with ROS as meta-operating system with following Features:
​
• Emotion detection based on User Facial landmark’s input (Mediapipe model).
• Emotion Display Expression on 2.8" TFT LCD emdedded on UNO.
• Two mode of locomotion: (i) Crawling (ii) DC motor based.
• Mycroft based Voice assistance.
OTHER PROJECTS
1] Gestural Remotely Controlled 3-DOF Manipulator with Hand-Pose Detection
• TFLite model used for hand landmark pose estimation
• Euclidean distance change leads to change in dutycycle of respective servo motors.
2] Transfer Learning Based Early Fire Detection System
• Early fire detection system based on deep learning and OpenCV, with customized InceptionV3 and CNN architectures for both interior and outdoor detection.
• OpenCV was used for live detection on a webcam with 980 photos for training and 239 images.
• For validation, with a training accuracy of 98.04 and a validation accuracy of 96.43
3] Wearable Palm Sized Device with Real-Time I2T Conversion
• Real-time Image to text conversion along with 20+ target translation languages with modell accuracy of 86.23 percent.
• Our device also convert those processed/translated text into audio data (in the development phase)