Researchers from the College of Information Technology Publish an International Study on Mesh Structure Generation Using GANs and SIFT Features

Duhaa Fadill Abbas
In a significant scientific achievement added to the research and development record of the Software Department at the College of Information Technology, doctoral student Raed Abdulreda Shekhan, under the supervision of Prof. Dr. Tawfiq Abdulkhaliq Al-Asadi, has published a peer-reviewed scientific paper titled:

"Generation of Fake Mesh Structures Using GANs with SIFT-Based Feature Preservation"

The study was published in the Journal of Information Systems Engineering and Management, a reputable international, peer-reviewed journal specializing in information systems, software engineering, and artificial intelligence applications.

Research Summary:
This study introduces an innovative approach that employs Generative Adversarial Networks (GANs), incorporating a fully independent layer architecture for both the generator and discriminator networks. As an application of deep learning techniques, this method allows the model to effectively capture intricate patterns within real-world data while mitigating the issue of overfitting when working with vector-based data representations, as opposed to traditional raster image data.

The core idea lies in training the GAN using vectorized data while preserving features extracted through the Scale-Invariant Feature Transform (SIFT) algorithm. This enhances the accuracy and quality of the generated results and opens new avenues for advanced image processing and complex data structure modeling.

This publication highlights the College of Information Technology s ongoing commitment to promoting high-impact scientific research and fostering innovation in artificial intelligence and data analysis.

Access the full article:
https://www.jisem-journal.com/index.php/journal/article/view/6611