This work introduces MonuSegFormer, a heritage-specific hybrid architecture coupling a pretrained Swin-B encoder with a Multi-Scale Atrous Fusion module and a CBAM-augmented progressive decoder to address severe class imbalance in Moroccan historical monuments.
Abstract
Semantic segmentation of architectural elements in cultural heritage sites lies at the intersection of computer vision and digital preservation. Moroccan historical monuments spanning mosques, madrasas, royal gates (babs), and mausoleums across Fez, Rabat, Marrakech, Meknes, and Tetouan present unique challenges, including extreme texture ambiguity between weathered wall surfaces and background, pronounced multi-scale variation from individual window and door openings to full rooftop surfaces spanning tens of metres, and severe class imbalance. We introduce MonuSegFormer, a heritage-specific hybrid architecture coupling a pretrained Swin-B encoder (ImageNet-22K) with a Multi-Scale Atrous Fusion (MSAF) module and a CBAM-augmented progressive decoder. Evaluated on the Moroccan Monuments Dataset (MMD, 2686 annotated RGB images, 5 semantic classes, 5 cities) on the held-out test set (403 images), MonuSegFormer achieves mIoU = 81.7%, outperforming SegFormer-B5 (76.4%), Mask2Former (78.9%), and DeepLabV3+ (71.3%). A composite CE+Dice+Focal loss with class-balanced weights addresses severe class imbalance, yielding the largest gains on the minority classes Door (+5.3 p.p.) and Window (+5.3 p.p.) over the next-best Mask2Former. All qualitative results are produced by real trained-model inference on held-out test images.
Objectives. We propose a modern method for semantic segmentation of ultra-high-resolution (4K) video frames in the field of Earth remote sensing using a modification of the DeepLabV3+ convolutional neural network.
Methods. The method is aimed at solving two critical problems: the limited receptive field of the model w...
This work proposes CMAE-UNet, a camouflage suppression and edge enhancement model based on the UNet architecture that robustly withstands field interference, providing strong technological support for automated tomato detection, intelligent harvesting, and growth monitoring.
Jian-Hua Zheng, Huang-Hui Zhao, Xiaoshan Ma et al.· AgriEngineering· 0 citations
This study proposes MLLAForestCD, a three-class pixel-level semantic change-detection network for Eucalyptus plantations, which uses an MLLA encoder to model the long-range spatial context with efficient linear attention, while a dual-axis change extractor reorganizes paired bi-temporal features through complementary l...
Experimental results on the ISPRS Potsdam and Vaihingen benchmark datasets show that the proposed CNN-Transformer model attains highly competitive segmentation performance while maintaining computational efficiency.
Wambugu Naftaly, Rui-Sheng Wang, Bo Guo et al.· Modeling Earth Systems and E...· 0 citations
This work proposes ViCo-SAM3, a Vision-Conditioned alignment framework designed for OVCOS, which introduces vision-conditioned (ViCo) module, which dynamically modulates text embeddings with global visual context, enabling textual representations to adapt to the current image content and thereby effectively bridges the...
Qiang-Qiang Zhou, Wengang Tang, Yong Chen et al.· 1 citation
EpigraphNet, a segmentation-guided transformer pipeline evaluated on the Persepolis Fortification Archive, is introduced, and a lower Spearman correlation between sign frequency and per-class performance indicates more balanced recognition across frequent and rare classes.
Utsav Poudel, Rasik Bhattarai, Siddhartha Pathak et al.· 0 citations
We use cookies to run the site and, with your consent, for analytics and to show ads.
See our Cookie Policy.