Galana Fekadu Asafa

Mr. Galana Fekadu Asafa

Computer Science and Technology, Central South University, Changsha, Guangdong, China
orcid

Research Focus & Metrics

Academic Career Length

2Years (2025-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
17 April 2025 in Electronics
Verified
Journal Article
Open access
Sheikh Sohan MamunSheikh Sohan Mamun
Sheikh Sohan Mamun
SRShengbing Ren
Shengbing Ren
MD Youshuf Khan RakibMD Youshuf Khan Rakib
MD Youshuf Khan Rakib
Galana Fekadu AsafaGalana Fekadu Asafa
Galana Fekadu Asafa
Multi-view 3D reconstruction aims to discover 3D characteristics based on visual information captured across multiple viewpoints. Transformer networks have shown remarkable success in various computer vision tasks, including multi-view 3D reconstruction. However, the reconstruction of accurate 3D shapes faces challenges when trying to efficiently extract and merge features across views. The existing frameworks struggled to capture the subtle relationships between the views, resulting in a poor reconstruction. To address this issue, we present a new framework, WGA-SWIN, for 3D reconstruction using multi-view objects. Our method introduces a Window Grouping Attention (WGA) mechanism that uses group tokens from different views for each window attention operation, enabling efficient inter-view and intra-view feature extraction. Diversity among various groups in a model contributes to the richness of feature learning, which results in advanced and dependable feature learning, resulting in more comprehensive and robust representations. Within the encoder swin transformer blocks, we integrated WGA to utilize both hierarchical design and shifted window attention mechanisms for efficient multi-view feature extraction. In addition, we developed a progressive hierarchical decoder that combines swin transformer blocks with 3D convolutions to utilize voxel representation, resulting in a high resolution for obtaining high-quality 3D reconstructions with fine structural details. The experimental results on the benchmark datasets ShapeNet and Pix3D demonstrate that our work achieves state-of-the-art (SOTA) performance, outperforming existing methods in both single-view and multi-view 3D reconstruction, beyond the capabilities of current technologies. We lead by 0.95% and 1.07% in both IoU and F-Scores respectively, which demonstrates the robustness of our method.

12 March 2025 in Electronics
Verified
Journal Article
Open access
PDF file
Galana Fekadu AsafaGalana Fekadu Asafa
Galana Fekadu Asafa
SRShengbing Ren
Shengbing Ren
Sheikh Sohan MamunSheikh Sohan Mamun
Sheikh Sohan Mamun
KGKaleb Amsalu Gobena
Kaleb Amsalu Gobena
Reconstructing 3D objects from single-view images has acquired significant interest due to their wide-ranging applications in robotics, autonomous vehicles, virtual reality, and augmented reality. Current methods, including voxel-based and point cloud-based approaches, face critical challenges such as irregular point distributions and an inability to preserve complex object details, which result in suboptimal reconstructions. To address these limitations, we propose DepthCloud2Point, a framework that combines depth maps, image features, and an initial point cloud to generate detailed and accurate 3D point clouds. Depth maps are employed to provide rich spatial cues, resolving depth ambiguities, while the initial point cloud serves as a geometric prior to ensure uniform point distribution. These components are integrated into a unified pipeline, where the encoder extracts semantic and geometric features, and the generator synthesizes high-fidelity 3D reconstructions. Our approach is trained end-to-end on both synthetic and real-world datasets, achieving state-of-the-art performance. Quantitative results on the ShapeNet dataset show that DepthCloud2Point outperforms 3D-LMNet by 19.07% in CD and 38.86% in EMD, and Pixel2Point by 18.77% in CD and 19.25% in EMD. We also perform a qualitative study that shows that our approach is able to generate reconstructions that closely align with ground truth, capturing intricate object details and maintaining spatial coherence, confirming its outperforming over the 3D-LMNet and Pixel2Point.

30 April 2026 in PLOS ONE
Verified
Journal Article
Open access
PDF file
KGKaleb Amsalu Gobena
Kaleb Amsalu Gobena
MD Youshuf Khan RakibMD Youshuf Khan Rakib
MD Youshuf Khan Rakib
FTFiseha Berhanu Tesema
Fiseha Berhanu Tesema
Galana Fekadu AsafaGalana Fekadu Asafa
Galana Fekadu Asafa
SRShengbing Ren
Shengbing Ren
Infrastructure cracks are critical indicators of structural deterioration in pavements, bridges, and buildings. Automated crack segmentation has therefore become an important component of structural health monitoring systems. However, accurate pixel-level crack segmentation on resource-constrained devices remains challenging due to the thin, low-contrast, and curvilinear morphology of cracks, as well as severe foreground–background class imbalance. To address these challenges, we propose LiteCrackSeg, a lightweight hybrid CNN–transformer architecture designed for efficient and accurate crack segmentation. The proposed framework adopts a hybrid MobileViT encoder that captures both local spatial details and long-range contextual dependencies while maintaining a compact model size. To enhance morphological sensitivity to elongated crack structures, we introduce a Morphology-Aware MobileViT (MAM-ViT) bottleneck, which integrates dual-branch Dynamic Snake Convolutions (DSConv) to align receptive fields with crack trajectories. Furthermore, a transformer-based decoder with local self-attention progressively reconstructs spatial details, while an attention-guided multi-scale fusion strategy improves boundary precision and structural continuity. To mitigate severe class imbalance, the model is trained using the Tversky loss, which explicitly balances false positives and false negatives. Extensive experiments on three public crack segmentation datasets (DeepCrack, CrackMap, and TUT) demonstrate that LiteCrackSeg achieves state-of-the-art segmentation performance while maintaining high computational efficiency. The proposed model requires only 2.72M parameters and 3.23 GFLOPs, achieving real-time inference at 56 FPS on 512 × 512 images, making it suitable for deployment on resource-constrained edge devices for practical infrastructure inspection.