Diego Resende Faria, Ph.D.
      Senior Lecturer in Computer Science, Aston University, Birmingham, UK


  scroll with the mouse to navigate and select an option at the main menu below




Affect Recognition through Facial Expressions using Dyanmic Bayesian Mixture Model
1 - tests with participants watching emotional ads (stimulated emotion)
2 - Tests using the humanoid robot NAO for human-robot interaction.


An EMG-controlled Low-cost Prosthetic Hand For Academia
A low-cost (5 DoF) 3D printed robotic hand controlled by electromyography (EMG).   Using deep learning techniques (LSTM) to train and classify simple gesture primitives, such as open and close hand. Work under progress. As extension, we are using Myo Gesture Control Armband (with 8 EMG channels) to recognise  more complex gestures.

Child-Robot Interaction (in collaboration with University of Coimbra, Portugal): Tests using the NAO robot to engage children during interaction. The communication is explored through gestural, verbal and physical interaction. Interdisciplinary Work (robotics / psychology), partnership between ISR-UC, CINEICC and APCC, Portugal. The objective was to evaluate the acceptance of children and parents regarding the use of robotics  technology in different scenarios (pediatric and educational settings).  

Probabilistic Human Daily Activity Recognition towards-Robot-assisted Living part I: The main idea is to use the DBMM framework in order to recognise human daily activities for natural human-robot interaction, herein, towards indoor monitoring of humans (e.g. elderly) in their daily routine and to detect possible risk situations. Once a risk situation is detected, the mobile robot must be able to react accordingly, e.g., if the person is lying down on the floor, the robot will ask if this person needs any help, and if the answer is "yes", then the robot will call (e.g. by skype) a relative or a medical doctor, sending a picture of the situation. Publications related to it: [17] [18] [19] in the page of publication list. Robot Programming (in ROS) with Mario Vieira (MSc. Student)


Probabilistic Human Daily Activity Recognition towards-Robot-assisted Living part II: This video shows an application useful for robot-assisted living where a robot is able to monitor the human daily activity. The robot also anticipates possible collision between human and robot by predicting when the human trajectory is coming towards the robot location. All technical details on implementation using ROS (navigation/SLAM, activity recognition, robot reaction) can be found in:
- Mario Vieira, Diego R. Faria, Urbano Nunes, "Real-time Application for Monitoring Human Daily Activities and Risk Situations in Robot-assisted Living". Proceedings of Robot'15: 2nd Iberian Robotics Conference. Lisbon, Portugal 2015.

   Social Activity Recognition Dataset: coming soon (RGB-D data will be publicly available). Work in Collaboration between ISR-UC and L-CAS, University of Lincoln, Uk. See publication list for more details.
Simulation of robot grasping for the dexterous shadow hand: Testing possible grasp types for pick-up and place given an object point cloud. In-hand exploration of objects using Probabilistic Volumetric Map:  3D Object shape representation through kinesthetic information. More details at Publications List: Journal [1] Conf.[11] [8]. Youtube video; https://youtu.be/GFWGVB21dXo
Object Identification through Hand Configurations (Grasp Types) and shape representation by in-hand Exploration. More details can be found at Publications List: Conf. [16] and Ph.D. Thesis. Youtube video: https://youtu.be/pNqheIg3aLk