UFAZ Teacher Nigar Alishzada Presents Research at International Congress

Nigar Alishzada, a teacher at the French-Azerbaijani University (UFAZ), recently participated in the 10th World Congress on Electrical Engineering and Computer Systems and Science, held from August 19-21 in Barcelona, Spain. At this prestigious event, Alishzada presented her research paper titled “Transfer Learning with Inflated 3D CNN for Word-Level Recognition for Azerbaijani Sign Language.”

Her research focuses on sign language recognition (SLR), a critical area of study for enhancing communication for the Deaf and Hard of Hearing (DHH) community. Sign language recognition aims to automate the understanding of sign language, treating each sign as a distinct class. While significant progress has been made using deep learning techniques, many sign languages, including Azerbaijani Sign Language, still lack the extensive datasets needed to train effective models.

To address this challenge, Alishzada's study explores the use of transfer learning, a method that applies knowledge from a well-established task with ample data to improve performance on a task with limited data. By leveraging an inflated 3D convolutional neural network (CNN) architecture, her research demonstrates how transfer learning can enhance SLR for Azerbaijani Sign Language. The approach involves pre-training a network on a related task and then fine-tuning it on a smaller dataset specific to Azerbaijani Sign Language, showing promising results in improving recognition accuracy.

Alishzada's work represents a significant advancement in the field of SLR, offering new possibilities for improving communication tools for the DHH community in Azerbaijan.

For those interested in exploring this important research further, the full article is available at this link: https://avestia.com/EECSS2024_Proceedings/files/paper/MVML/MVML_106.pdf