In recent years, advances in computer hardware, graphics rendering algorithms and computer vision have enabled the utilization of 3D building reconstructions in the fields of archeological structure restoration and urban planning. This paper deals with the reconstruction of realistic 3D models of buildings façades, in the urban environment for cultural heritage. The proposed approach is an extension of our previous work in this research topic, which introduced a methodology for accurate 3D realistic façade reconstruction by defining and exploiting a relation between stereoscopic image and tacheometry data. In this work, we re-purpose well known deep neural network architectures in the fields of image segmentation and single image depth prediction, for the tasks of façade structural element detection, depth point-cloud generation and protrusion estimation, with the goal of alleviating drawbacks in our previous design, resulting in a more light-weight, robust, flexible and cost-effective design.

3D building façade reconstruction has become a very popular topic in various applications related to restoration and preservation of architectural structures as well as urban planning. This paper deals with the reconstruction of realistic 3D models of buildings façades, in the urban environment for cultural heritage. We present an approach that enables the relation of stereoscopic images with tacheometry data. The proposed multimodal fusing scheme results in an accurate 3D realistic façade reconstruction and provides a fast and low cost solution. In the first stage of the proposed approach a 2D skeleton of the building is extracted from the viewed scene using Active Contour and Hough line extraction. The next stage of our method utilizes depth information, extracted from a stereoscopic layout, to infer the structural details of inner façade structures, such as windows or doors. In the final stage, the structural information extracted from the image data is integrated with georeferenced point datasets. The final output of our method is a georeferenced 3D model of the structure’s façade, which can be further refined with the use of image-driven texture information.

Accurate flood detection is critical for disaster response, yet the scarcity of diverse annotated datasets hinders robust model development. Existing resources typically suffer from limited geographic scope and insufficient annotation granularity, restricting the generalization capabilities of computer vision methods. To bridge this gap, we introduce AIFloodSense, a comprehensive evaluation benchmark designed to advance domain-generalized Artificial Intelligence for climate resilience. The dataset comprises 470 high-resolution aerial images capturing 230 distinct flood events across 64 countries and six continents. Unlike prior benchmarks, AIFloodSense ensures exceptional global diversity and temporal relevance (2022–2024), supporting three complementary tasks: (i) Image Classification, featuring novel sub-tasks for environment type, camera angle, and continent recognition; (ii) Semantic Segmentation, providing precise pixel-level masks for flood, sky, buildings, and background; and (iii) Visual Question Answering (VQA), enabling natural language reasoning for disaster assessment. We provide baseline benchmarks for all tasks using state-of-the-art architectures, demonstrating the dataset’s complexity and its utility in fostering robust AI tools for environmental monitoring. Crucially, we show that despite its compact size, AIFloodSense enables better generalization on external test sets than much larger alternatives, validating the premise that rigorous diversity is more effective than scale for training robust flood detection models, and is made publicly available to accelerate further research in the field.

In the field of human action recognition (HAR), the recognition of actions with large duration is hindered by the memorization capacity limitations of the standard probabilistic and recurrent neural network (R-NN) approaches that are used for temporal sequence modeling. The simplest remedy is to employ methods that reduce the input sequence length, by performing window sampling, pooling, or key-frame extraction. However, due to the nature of the frame selection criteria or the employed pooling operations, the majority of these approaches do not guarantee that the useful, discriminative information is preserved. In this work, we focus on the case of repetitive actions. In such actions, a discriminative, core execution motif is maintained throughout each repetition, with slight variations in execution style and duration. Additionally, scene appearance may change as a consequence of the action. We exploit those two key observations on the nature of repetitive actions to build a compact and efficient representation of long actions by maintaining the discriminative sample information and removing redundant information which is due to task repetitiveness. We show that by partitioning an input sequence based on repetition and by treating each repetition as a discrete sample, HAR models can achieve an increase of up to 4% in action recognition accuracy. Additionally, we investigate the relation between the dataset and action set attributes with this strategy and explore the conditions under which the utilization of repetitiveness for input sequence sampling, is a useful preprocessing step in HAR. Finally, we suggest deep NN design directions that enable the effective exploitation of the distinctive action-related information found in repetitiveness, and evaluate them with a simple deep architecture that follows these principles.

We present a novel methodology for generating and filtering synthetic Unmanned Aerial Vehicle (UAV) flood imagery to enhance the generalization capabilities of segmentation models. Our framework combines text-to-image synthesis and image inpainting, using curated prompts and real-world segmentation masks to produce diverse and realistic flood scenes. To overcome the lack of human annotations, we employ an unsupervised pseudo-labeling method that generates segmentation masks based on floodwater appearance characteristics. We further introduce a filtering stage based on outlier detection in feature space to improve the realism of the synthetic dataset. Experimental results on five state-of-the-art flood segmentation models show that synthetic data can closely match real data in training performance, and combining both sources improves model robustness by 1–7%. Finally, we investigate the impact of prompt design on the visual fidelity of generated images and provide qualitative and quantitative evidence of distributional similarity between real and synthetic data.