Authors:
Rahma Aloui
1
;
Pranav Martini
1
;
Pandu Devarakota
2
;
Apurva Gala
2
and
Shishir Shah
1
Affiliations:
1
University of Houston, Houston, TX, U.S.A.
;
2
Shell Information Technology International Inc., Houston, TX, U.S.A.
Keyword(s):
Image Segmentation, UNet Variants, Gabor Filters, Spatial-Channel Squeeze-and-Excitation, Multi-Scale Feature Fusion, Gabor Convolution, Retinal Vessels Images, Seismic Images.
Abstract:
Accurate delineation of critical features, such as salt boundaries in seismic imaging and fine structures in medical images, is essential for effective analysis and decision-making. Traditional convolutional neural networks (CNNs) often face difficulties in handling complex data due to variations in scale, orientation, and noise. These limitations become particularly evident during the transition from proof-of-concept to real-world deployment, where models must perform consistently under diverse conditions. To address these challenges, we propose GAM-UNet, an advanced segmentation architecture that integrates learnable Gabor filters for enhanced edge detection, SCSE blocks for feature refinement, and multi-scale fusion within the U-Net framework. This approach improves feature extraction across varying scales and orientations. Trained using a combined Binary Cross-Entropy and Dice loss function, GAM-UNet demonstrates superior segmentation accuracy and continuity, outperforming existi
ng U-Net variants across diverse datasets.
(More)