Pre-Dissertation Seminar by Ph.D. Candidate Safaa Hadi Khudair Abis in the Department of Software

Duhaa Fadill Abbas
As part of the ongoing scientific activities at the Department of Software – College of Information Technology, University of Babylon, Ph.D. candidate Safaa Hadi Khudair Abis delivered his pre-dissertation seminar on his proposed research titled: "Land Cover Classification of Synthetic Aperture Radar Images Using Machine Learning", under the supervision of Prof. Dr. Hidhab Khalid Abis, and in the presence of the department’s scientific committee.

In his presentation, the researcher emphasized the growing significance of deep learning techniques in the field of remote sensing, particularly for land cover classification. He explained that the effectiveness of these methods relies heavily on the availability of large volumes of high-quality training samples—an essential yet challenging factor in achieving accurate results.

The dissertation proposes an innovative strategy to enhance and expand the temporal sequences of training samples. This approach involves utilizing caching statistics and band selection to increase data volume and reduce class overlap within buffer zones. The study further integrates an attention mechanism to fuse spatial and temporal feature vectors extracted from multiple LSTM cells, in conjunction with a U-Net architecture—an approach that significantly improves classification precision.