Large foundation models have emerged in the last years and are pushing performance boundaries for a variety of tasks. Training or even finetuning such models demands vast datasets and computational resources, which are often scarce and costly. Adaptation methods provide a computationally efficient solution to address these limitations by allowing such models to be finetuned on small amounts of data and computing power. This is achieved by appending new trainable modules to frozen backbones with only a fraction of the trainable parameters and fitting only these modules on novel tasks. Recently, the VeRA adapter was shown to excel in parameter-efficient adaptations by utilizing a pair of frozen random low-rank matrices shared across all layers. In this paper, we propose PVeRA, a probabilistic version of the VeRA adapter, which modifies the low-rank matrices of VeRA in a probabilistic manner. This modification naturally allows handling inherent ambiguities in the input and allows for different sampling configurations during training and testing. A comprehensive evaluation was performed on the VTAB-1k benchmark and seven adapters, with PVeRA outperforming VeRA and other adapters. Our code for training models with PVeRA and benchmarking all adapters is available here.

Multiple Instance Learning (MIL) offers a natural solution for settings where only coarse, bag-level labels are available, without having access to instance-level annotations. This is usually the case in digital pathology, which consists of gigapixel-sized images. While deterministic attention-based MIL approaches achieve strong bag-level performance, they often overlook the uncertainty inherent in instance relevance. In this paper, we address the lack of uncertainty quantification in instance-level attention scores by introducing SGPMIL, a new probabilistic attention-based MIL framework grounded in Sparse Gaussian Processes (SGP). By learning a posterior distribution over attention scores, SGPMIL enables principled uncertainty estimation, resulting in more reliable and calibrated instance relevance maps. Our approach not only preserves competitive bag-level performance but also significantly improves the quality and interpretability of instance-level predictions under uncertainty. SGPMIL extends prior work by introducing feature scaling in the SGP predictive mean function, leading to faster training, improved efficiency, and enhanced instance-level performance. Extensive experiments on multiple well-established digital pathology datasets highlight the effectiveness of our approach across both bag- and instance-level evaluations. Our code is available at https://github.com/mandlos/SGPMIL.

Accurate detection of clinically significant prostate cancer (csPCa) remains a major challenge in prostate MRI interpretation. Radiomics provides a noninvasive approach by extracting quantitative descriptors from imaging data, such as Apparent Diffusion Coefficient (ADC) maps. In this study, we systematically benchmarked radiomics-based machine learning pipelines across two validation scenarios: (1) repeated cross-validation on a multicenter dataset, and (2) nested cross-validation with external testing on the public PROSTATEx dataset. We evaluated eight feature selection methods, fifteen classifiers, and the impact of ComBat harmonization. The best-performing pipeline-Recursive Feature Elimination (RFE) with 20 features and a Random Forest classifier, without ComBat-achieved an internal AUCPR of$0.880 \pm 0.061$and an external AUC-PR of 0.717, with an F1-score of 0.748. Although ComBat improved calibration (F1score) in some cases, its effect on external discrimination was limited. Notably, we found that selecting 20 features yielded optimal generalization performance, supporting the “rule of 10 ” guideline for feature-to-sample ratio. These findings highlight the importance of rigorous feature selection and classifier design in developing robust and generalizable radiomics models for csPCa detection.

Reliable risk assessment for carotid atheromatous disease requires integrating diverse clinical and imaging information in a transparent and interpretable manner. This study investigates state-of-the-art large vision-language models (LVLMs) for multimodal carotid plaque assessment by integrating ultrasound imaging with structured clinical, demographic, laboratory, and biomarker data. A framework that simulates realistic diagnostic scenarios through interview-style question sequences is proposed, comparing open-source LVLMs including general-purpose and medically tuned models. Zero-shot experiments reveal that while most LVLMs accurately identify imaging modality and anatomy, all perform poorly in risk stratification. To address this, LLaVa-NeXT-Vicuna is adapted using low-rank adaptation (LoRA), achieving substantial improvements in stroke risk stratification. Integrating multimodal tabular data as text further enhances specificity and balanced accuracy, yielding competitive performance compared to prior CNN baselines. Our findings highlight both promise and limitations of LVLMs in ultrasound-based cardiovascular risk prediction, underscoring the importance of multimodal integration, domain adaptation, and model calibration for clinical translation.

Whole slide image (WSI) analysis in digital pathology presents unique challenges due to the gigapixel resolution of WSIs and the scarcity of dense supervision signals. While Multiple Instance Learning (MIL) is a natural fit for slide-level tasks, training robust models requires large and diverse datasets. Even though image augmentation techniques could be utilized to increase data variability and reduce overfitting, implementing them effectively is not a trivial task. Traditional patch-level augmentation is prohibitively expensive due to the large number of patches extracted from each WSI, and existing feature-level augmentation methods lack control over transformation semantics. We introduce HistAug, a fast and efficient generative model for controllable augmentations in the latent space for digital pathology. By conditioning on explicit patch-level transformations (e.g., hue, erosion), HistAug generates realistic augmented embeddings while preserving initial semantic information. Our method allows the processing of a large number of patches in a single forward pass efficiently, while at the same time consistently improving MIL model performance. Experiments across multiple slide-level tasks and diverse organs show that HistAug outperforms existing methods, particularly in low-data regimes. Ablation studies confirm the benefits of learned transformations over noise-based perturbations and highlight the importance of uniform WSI-wise augmentation. Code is available at https://github.com/MICS-Lab/HistAug.

Progress in a research field can be hard to assess, in particular when many concurrent methods are proposed in a short period of time. This is the case in digital pathology, where many foundation models have been released recently to serve as feature extractors for tile-level images, being used in a variety of downstream tasks, both for tile- and slide-level problems. Benchmarking available methods then becomes paramount to get a clearer view of the research landscape. In particular, in critical domains such as healthcare, a benchmark should not only focus on evaluating downstream performance, but also provide insights about the main differences between methods, and importantly, further consider uncertainty and robustness to ensure a reliable usage of proposed models. For these reasons, we introduce THUNDER, a tile-level benchmark for digital pathology foundation models, allowing for efficient comparison of many models on diverse datasets with a series of downstream tasks, studying their feature spaces and assessing the robustness and uncertainty of predictions informed by their embeddings. THUNDER is a fast, easy-to-use, dynamic benchmark that can already support a large variety of state-of-the-art foundation, as well as local user-defined models for direct tile-based comparison. In this paper, we provide a comprehensive comparison of 23 foundation models on 16 different datasets covering diverse tasks, feature analysis, and robustness. The code for THUNDER is publicly available at https://github.com/MICS-Lab/thunder.

Characterizing cancer presents a delicate challenge as it involves deciphering complex biological interactions within the tumor's microenvironment. Clinical trials often provide histology images and molecular profiling of tumors, which can help understand these interactions. Despite recent advances in representing multimodal data for weakly supervised tasks in the medical domain, achieving a coherent and interpretable fusion of whole slide images and multi-omics data is still a challenge. Each modality operates at distinct biological levels, introducing substantial correlations between and within data sources. In response to these challenges, we propose a novel deep-learning-based approach designed to represent multi-omics & histopathology data for precision medicine in a readily interpretable manner. While our approach demonstrates superior performance compared to state-of-the-art methods across multiple test cases, it also deals with incomplete and missing data in a robust manner. It extracts various scores characterizing the activity of each modality and their interactions at the pathway and gene levels. The strength of our method lies in its capacity to unravel pathway activation through multimodal relationships and to extend enrichment analysis to spatial data for supervised tasks. We showcase its predictive capacity and interpretation scores by extensively exploring multiple TCGA datasets and validation cohorts. The method opens new perspectives in understanding the complex relationships between multimodal pathological genomic data in different cancer types and is publicly available on Github.

Representation learning methods that discover task and/or data-specific characteristics are very popular for a variety of applications. However, their application to 3D medical images is restricted by the computational cost and their inherent subtle differences in intensities and appearance. In this paper, a novel representation learning scheme for extracting representations capable of distinguishing high-uptake regions from 3D 18F-Fluorodeoxyglucose positron emission tomography (FDG-PET) images is proposed. In particular, we propose a novel position-enhanced learning scheme effectively incorporating semantic and position-based features through our proposed Position Encoding Block (PEB) to produce highly informative representations. Such representations incorporate both semantic and position-aware features from high-dimensional medical data, leading to general representations with better performance on clinical tasks. To evaluate our method, we conducted experiments on the challenging task of classifying high-uptake regions as either non-tumor or tumor lesions in Metastatic Melanoma (MM). MM is a type of cancer characterized by its rapid spread to various body sites, which leads to low survival rates. Extensive experiments on an in-house and a public dataset of whole-body FDG-PET images indicated an increase of 10.50% in sensitivity and 4.89% in F1-score against the baseline representation learning scheme while also outperforming state-of-the-art methods for classifying MM regions of interest. The source code will be available at https://github.com/theoVag/Representation-Learning-Sem-Pos.

Introducing interpretability and reasoning into Multiple Instance Learning (MIL) methods for Whole Slide Image (WSI) analysis is challenging, given the complexity of gigapixel slides. Traditionally, MIL interpretability is limited to identifying salient regions deemed pertinent for downstream tasks, offering little insight to the end-user (pathologist) regarding the rationale behind these selections. To address this, we propose Self-Interpretable MIL (SI-MIL), a method intrinsically designed for interpretability from the very outset. SI-MIL employs a deep MIL framework to guide an interpretable branch grounded on handcrafted pathological features, facilitating linear predictions. Beyond identifying salient regions, SI-MIL uniquely provides feature-level interpretations rooted in pathological insights for WSIs. Notably, SI-MIL, with its linear prediction constraints, challenges the prevalent myth of an inevitable trade-off between model interpretability and performance, demonstrating competitive results compared to state-of-the-art methods on WSI-level prediction tasks across three cancer types. In addition, we thoroughly benchmark the local-and global-interpretability of SI-MIL in terms of statistical analysis, a domain expert study, and desiderata of interpretability, namely, user-friendliness and faithfulness.

In Self-Supervised Learning (SSL), models are typically pretrained, fine-tuned, and evaluated on the same domains. However, they tend to perform poorly when evaluated on unseen domains, a challenge that Unsupervised Domain Generalization (UDG) seeks to address. Current UDG methods rely on domain labels, which are often challenging to collect, and domain-specific architectures that lack scalability when confronted with numerous domains, making the current methodology impractical and rigid. Inspired by contrastive-based UDG methods that mitigate spurious correlations by restricting comparisons to examples from the same domain, we hypothesize that eliminating style variability within a batch could provide a more convenient and flexible way to reduce spurious correlations without requiring domain labels. To verify this hypothesis, we introduce Batch Styles Standardization (BSS), a relatively simple yet powerful Fourier-based method to standardize the style of images in a batch specifically designed for integration with SSL methods to tackle UDG. Combining BSS with existing SSL methods offers serious advantages over prior UDG methods: (1) It eliminates the need for domain labels or domain-specific network components to enhance domain-invariance in SSL representations, and (2) offers flexibility as BSS can be seamlessly integrated with diverse contrastive-based but also non-contrastive-based SSL methods. Experiments on several UDG datasets demonstrate that it significantly improves downstream task performances on unseen domains, often outperforming or rivaling UDG methods. Finally, this work clarifies the underlying mechanisms contributing to BSS's effectiveness in improving domain-invariance in SSL representations and performances on unseen domains. Implementations of the extended SSL methods and BSS are provided at this [url](https://gitlab.com/vitadx/articles/towards-domain-invariant-ssl-through-bss).

Self-Supervised learning (SSL) with Joint-Embedding Architectures (JEA) has led to outstanding performances. All instantiations of this paradigm were trained using strong and well-established hand-crafted data augmentations, leading to the general belief that they are required for the proper training and performance of such models. On the other hand, generative reconstruction-based models such as BEIT and MAE or Joint-Embedding Predictive Architectures such as I-JEPA have shown strong performance without using data augmentations except masking. In this work, we challenge the importance of invariance and data-augmentation in JEAs at scale. By running a case-study on a recent SSL foundation model -- DINOv2 -- we show that strong image representations can be obtained with JEAs and only cropping without resizing provided the training data is large enough, reaching state-of-the-art results and using the least amount of augmentation in the literature. Through this study, we also discuss the impact of compute constraints on the outcomes of experimental deep learning research, showing that they can lead to very different conclusions.

Exploring and identifying a good feature representation to describe high-dimensional datasets is a challenge of prime importance. However, plenty of feature selection techniques and distance metrics exist, which entails an intricacy for identifying the one best suited to the task. This paper provides an algorithm to design high-order distance metrics over a sparse selection of features dedicated to classification. Our approach is based on Conditional Random Field (CRF) energy minimization and Dual Decomposition, which allow efficiency and great flexibility in the considered features. The optimization technique ensures the tractability of high-dimensionality problems using hundreds of features and samples. Our approach is evaluated on synthetic data as well as on Covid-19 patient stratification. Comparisons with state-of-the-art baselines and our proposed method on different classification results prove the learned metric’s relevance.

This paper presents an investigation into the zero-shot performance of pre-trained foundation models on remote sensing tasks. Recent advances in self-supervised learning suggest that these models, when trained on vast amounts of unsupervised data, could potentially improve generalization across a number of downstream tasks. Our study offers an empirical evaluation of these models on standard remote-sensing benchmarks such as EuroSAT and BigEarthNet-S2, with the intent to confirm whether these models have encountered satellite imagery during their training phase. Moreover, we examine the impact of adding a geospatial domain-specific textual description of classes, contrasting it with the standard class-based prompts. Our findings indicate that the fine-tuned BLIP models exhibit superior zero-shot performance on these benchmarks compared to their standard counterparts, signifying that fine-tuning on standard benchmarks enhances performance. Furthermore, the addition of geospatial context variably influences performance depending on the specific model and dataset. This work provides crucial insights into the applicability of foundation models in remote sensing tasks and lays the groundwork for further research.

The robustness of image segmentation has been an important research topic in the past few years as segmentation models have reached production-level accuracy. However, like classification models, segmentation models can be vulnerable to adversarial perturbations, which hinders their use in critical-decision systems like healthcare or autonomous driving. Recently, randomized smoothing has been proposed to certify segmentation predictions by adding Gaussian noise to the input to obtain theoretical guarantees. However, this method exhibits a trade-off between the amount of added noise and the level of certification achieved. In this paper, we address the problem of certifying segmentation prediction using a combination of randomized smoothing and diffusion models. Our experiments show that combining randomized smoothing and diffusion models significantly improves certified robustness, with results indicating a mean improvement of 21 points in accuracy compared to previous state-of-the-art methods on Pascal-Context and Cityscapes public datasets. Our method is independent of the selected segmentation model and does not need any additional specialized training procedure.

Image registration is a fundamental medical image analysis task, and a wide variety of approaches have been proposed. However, only a few studies have comprehensively compared medical image registration approaches on a wide range of clinically relevant tasks. This limits the development of registration methods, the adoption of research advances into practice, and a fair benchmark across competing approaches. The Learn2Reg challenge addresses these limitations by providing a multi-task medical image registration data set for comprehensive characterisation of deformable registration algorithms. A continuous evaluation will be possible at https://learn2reg.grand-challenge.org. Learn2Reg covers a wide range of anatomies (brain, abdomen, and thorax), modalities (ultrasound, CT, MR), availability of annotations, as well as intra- and inter-patient registration evaluation. We established an easily accessible framework for training and validation of 3D registration methods, which enabled the compilation of results of over 65 individual method submissions from more than 20 unique teams. We used a complementary set of metrics, including robustness, accuracy, plausibility, and runtime, enabling unique insight into the current state-of-the-art of medical image registration. This paper describes datasets, tasks, evaluation methods and results of the challenge, as well as results of further analysis of transferability to new datasets, the importance of label supervision, and resulting bias. While no single approach worked best across all tasks, many methodological aspects could be identified that push the performance of medical image registration to new state-of-the-art performance. Furthermore, we demystified the common belief that conventional registration methods have to be much slower than deep-learning-based methods.

During the last decades, the remote sensing community has gained access to a wide satellite imagery material, resulting in great progress on various applications. Most of these applications firstly require that the employed images are in the same coordinate system, without any registration errors that will deteriorate their performance. In this work we employ a multistep fully convolutional network in order to improve the image registration task. Moreover, we propose the relaxation of the registration constraints on the regions where changes have occurred. In this way, the model learns to keep the proper structures of objects that change in the multitemporal pair while at the same time provides dense deformations for the rest of the region. Our method is very efficient, fast and requires annotations for the regions of change only in training time. We conduct experiments on the very high resolution Attica VHR dataset comparing it with other deformable registration approaches from the literature. Our method outperforms all compared methods both quantitatively and qualitatively, setting up a foundation for further future research.

Precision medicine is a paradigm shift in healthcare relying heavily on genomics data. However, the complexity of biological interactions, the large number of genes as well as the lack of comparisons on the analysis of data, remain a tremendous bottleneck regarding clinical adoption. In this paper, we introduce a novel, automatic and unsupervised framework to discover low-dimensional gene biomarkers. Our method is based on the LP-Stability algorithm, a high dimensional center-based unsupervised clustering algorithm. It offers modularity as concerns metric functions and scalability, while being able to automatically determine the best number of clusters. Our evaluation includes both mathematical and biological criteria to define a quantitative metric. The recovered signature is applied to a variety of biological tasks, including screening of biological pathways and functions, and characterization relevance on tumor types and subtypes. Quantitative comparisons among different distance metrics, commonly used clustering methods and a referential gene signature used in the literature, confirm state of the art performance of our approach. In particular, our signature, based on 27 genes, reports at least 30 times better mathematical significance (average Dunn's Index) and 25% better biological significance (average Enrichment in Protein-Protein Interaction) than those produced by other referential clustering methods. Finally, our signature reports promising results on distinguishing immune inflammatory and immune desert tumors, while reporting a high balanced accuracy of 92% on tumor types classification and averaged balanced accuracy of 68% on tumor subtypes classification, which represents, respectively 7% and 9% higher performance compared to the referential signature.

Coronavirus disease 2019 (COVID-19) emerged in 2019 and disseminated around the world rapidly. Computed tomography (CT) imaging has been proven to be an important tool for screening, disease quantification and staging. The latter is of extreme importance for organizational anticipation (availability of intensive care unit beds, patient management planning) as well as to accelerate drug development through rapid, reproducible and quantified assessment of treatment response. Even if currently there are no specific guidelines for the staging of the patients, CT together with some clinical and biological biomarkers are used. In this study, we collected a multi-center cohort and we investigated the use of medical imaging and artificial intelligence for disease quantification, staging and outcome prediction. Our approach relies on automatic deep learning-based disease quantification using an ensemble of architectures, and a data-driven consensus for the staging and outcome prediction of the patients fusing imaging biomarkers with clinical and biological attributes. Highly promising results on multiple external/independent evaluation cohorts as well as comparisons with expert human readers demonstrate the potentials of our approach.

In this article, we present a deep multitask learning framework able to couple semantic segmentation and change detection using fully convolutional long short-term memory (LSTM) networks. In particular, we present a UNet-like architecture (L-UNet) that models the temporal relationship of spatial feature representations using integrated fully convolutional LSTM blocks on top of every encoding level. In this way, the network is able to capture the temporal relationship of spatial feature vectors in all encoding levels without the need to downsample or flatten them, forming an end-to-end trainable framework. Moreover, we further enrich the L-UNet architecture with an additional decoding branch that performs semantic segmentation on the available semantic categories that are presented in the different input dates, forming a multitask framework. Different loss quantities are also defined and combined together in a circular way to boost the overall performance. The developed methodology has been evaluated on three different data sets, i.e., the challenging bitemporal high-resolution Office National d'Etudes et de Recherches Aérospatiales (ONERA) Satellite Change Detection (OSCD) Sentinel-2 data set, the very high-resolution (VHR) multitemporal data set of the East Prefecture of Attica, Greece, and finally, the multitemporal VHR SpaceNet7 data set. Promising quantitative and qualitative results demonstrated that the synergy among the tasks can boost up the achieved performances. In particular, the proposed multitask framework contributed to a significant decrease in false-positive detections, with the F1 rate outperforming other state-of-the-art methods by at least 2.1% and 4.9% in the Attica VHR and SpaceNet7 data set cases, respectively. Our models and code can be found at https://github.com/mpapadomanolaki/multi-task-L-UNet.

We investigate the use of recent advances in deep learning and propose an end-to-end trainable multi-instance convolutional neural network within a mixture-of-experts formulation that combines information from two types of data-images and clinical attributes-for the diagnosis of lymphocytosis. The convolutional network learns to extract meaningful features from images of blood cells using an embedding level approach and aggregates them. Moreover, the mixture-of-experts model combines information from these images as well as clinical attributes to form an end-to-end trainable pipeline for diagnosis of lymphocytosis. Our results demonstrate that even the convolutional network by itself is able to discover meaningful associations between the images and the diagnosis, indicating the presence of important unexploited information in the images. The mixture-of-experts formulation is shown to be more robust while maintaining performance via. a repeatability study to assess the effect of variability in data acquisition on the predictions. The proposed methods are compared with different methods from literature based both on conventional handcrafted features and machine learning, and on recent deep learning models based on attention mechanisms. Our method reports a balanced accuracy of [Formula: see text] and outperfroms the handcrafted feature-based and attention-based approaches as well that of biologists which scored [Formula: see text], [Formula: see text] and [Formula: see text] respectively. These results give insights on the potentials of the applicability of the proposed method in clinical practice. Our code and datasets can be found at https://github.com/msahasrabudhe/lymphoMIL.

Semantic segmentation, especially for very high-resolution satellite data, is one of the pillar problems in the remote sensing community. Lately, deep learning techniques are the ones that set the state-of-the-art for a number of benchmark datasets, however, there are still a lot of challenges that need to be addressed, especially in the case of limited annotations. To this end, in this paper, we propose a novel framework based on deep neural networks that is able to address concurrently semantic segmentation and image reconstruction in an end to end training. Under the proposed formulation, the image reconstruction acts as a regularization, constraining efficiently the solution in the entire image domain. This self-supervised component helps significantly the generalization of the network for the semantic segmentation, especially in cases of a low number of annotations. Experimental results and the performed quantitative evaluation on the publicly available ISPRS (WGIII/4) dataset indicate the great potential of the developed approach.

The advent of multitemporal high resolution data, like the Copernicus Sentinel-2, has enhanced significantly the potential of monitoring the earth's surface and environmental dynamics. In this paper, we present a novel deep learning framework for urban change detection which combines state-of-the-art fully convolutional networks (similar to U-Net) for feature representation and powerful recurrent networks (such as LSTMs) for temporal modeling. We report our results on the recently publicly available bi-temporal Onera Satellite Change Detection (OSCD) Sentinel-2 dataset, enhancing the temporal information with additional images of the same region on different dates. Moreover, we evaluate the performance of the recurrent networks as well as the use of the additional dates on the unseen test-set using an ensemble cross-validation strategy. All the developed models during the validation phase have scored an overall accuracy of more than 95%, while the use of LSTMs and further temporal information, boost the F1 rate of the change class by an additional 1.5%.

Image registration in multimodal, multitemporal satellite imagery is one of the most important problems in remote sensing and essential for a number of other tasks such as change detection and image fusion. In this paper, inspired by the recent success of deep learning approaches we propose a novel convolutional neural network architecture that couples linear and deformable approaches for accurate alignment of remote sensing imagery. The proposed method is completely unsupervised, ensures smooth displacement fields and provides real time registration on a pair of images. We evaluate the performance of our method using a challenging multitemporal dataset of very high resolution satellite images and compare its performance with a state of the art elastic registration method based on graphical models. Both quantitative and qualitative results prove the high potentials of our method.

In this study, we propose a 3D deep neural network called U-ReSNet, a joint framework that can accurately register and segment medical volumes. The proposed network learns to automatically generate linear and elastic deformation models, trained by minimizing the mean square error and the local cross correlation similarity metrics. In parallel, a coupled architecture is integrated, seeking to provide segmentation maps for anatomies or tissue patterns using an additional decoder part trained with the dice coefficient metric. U-ReSNet is trained in an end to end fashion, while due to this joint optimization the generated network features are more informative leading to promising results compared to other deep learning-based methods existing in the literature. We evaluated the proposed architecture using the publicly available OASIS 3 dataset, measuring the dice coefficient metric for both registration and segmentation tasks. Our promising results indicate the potentials of our method which is composed from a convolutional architecture that is extremely simple and light in terms of parameters.

Semantic segmentation is currently a mainstream method for addressing several remote sensing applications, achieving recently remarkable performance by employing deep learning techniques. In particular, this is the case for pixel-wise dense classification models in very high resolution remote sensing datasets. In this paper, we exploit the use of a relatively deep architecture based on repetitive downscale-upscale processes that had been previously employed for human pose estimation tasks. By integrating such a model, we are aiming to capture and extract low-level details, such as small objects, object boundaries and edges. Experimental results and quantitative evaluation has been performed on the publicly available ISPRS (WGIII4) benchmark dataset indicating the potential of the proposed approach.

Automatic and accurate detection of man-made objects, such as buildings, is one of the main problems that the remote sensing community has been focusing on for the last decades. In this paper, we propose a Conditional Random Field (CRF) formulation which is using edge/boundary localization priors towards accurate building detection. These edge priors have been integrated/fused with the classification scores from a deep learning Convolutional Neural Network (CNN) architecture under a single energy formulation. The validation of the developed methodology had been performed on the recently published SpaceNet dataset. Experimental results and quantitative evaluation, based on different accuracy statistics, indicate the great potential of the proposed approach.

In this paper, a novel generic framework has been designed, developed and validated for addressing simultaneously the tasks of image registration, segmentation and change detection from multisensor, multiresolution, multitemporal satellite image pairs. Our approach models the inter-dependencies of variables through a higher order graph. The proposed formulation is modular with respect to the nature of images (various similarity metrics can be considered), the nature of deformations (arbitrary interpolation strategies), and the nature of segmentation likelihoods (various classification approaches can be employed). Inference of the proposed formulation is achieved through its mapping to an overparametrized pairwise graph which is then optimized using linear programming. Experimental results and the performed quantitative evaluation indicate the high potentials of the developed method.

The automated man-made object detection and building extraction from single satellite images is, still, one of the most challenging tasks for various urban planning and monitoring engineering applications. To this end, in this paper we propose an automated building detection framework from very high resolution remote sensing data based on deep convolutional neural networks. The core of the developed method is based on a supervised classification procedure employing a very large training dataset. An MRF model is then responsible for obtaining the optimal labels regarding the detection of scene buildings. The experimental results and the performed quantitative validation indicate the quite promising potentials of the developed approach.