Search

Vision foundation models have achieved remarkable progress across various image analysis tasks. In the image segmentation task, foundation models like the Segment Anything Model (SAM) enable generalizable zero-shot segmentation through user-provided prompts. However, SAM primarily trained on natural images, lacks the domain-specific expertise of medical imaging. This limitation poses challenges when applying SAM to medical image segmentation, including the need for extensive fine-tuning on specialized medical datasets and a dependency on manual prompts, which are both labor-intensive and require intervention from medical experts. This work introduces the Few-shot Adaptation of Training-frEe SAM (FATE-SAM), a novel method designed to adapt the advanced Segment Anything Model 2 (SAM2) for 3D medical image segmentation. FATE-SAM reassembles pre-trained modules of SAM2 to enable few-shot adaptation, leveraging a small number of support examples to capture anatomical knowledge and perform prompt-free segmentation, without requiring model fine-tuning. To handle the volumetric nature of medical images, we incorporate a Volumetric Consistency mechanism that enhances spatial coherence across 3D slices. We evaluate FATE-SAM on multiple medical imaging datasets and compare it with supervised learning methods, zero-shot SAM approaches, and fine-tuned medical SAM methods. Results show that FATE-SAM delivers robust and accurate segmentation while eliminating the need for large annotated datasets and expert intervention. FATE-SAM provides a practical, efficient solution for medical image segmentation, making it more accessible for clinical applications.

In the field of sensor fusion and state estimation for object detection and localization, ensuring accurate tracking in dynamic environments poses significant challenges. Traditional methods like the Kalman Filter (KF) often fail when measurements are intermittent, leading to rapid divergence in state estimations. To address this, we introduce SMART (Sensor Measurement Augmentation and Reacquisition Tracker), a novel approach that leverages high-frequency state estimates from the KF to guide the search for new measurements, maintaining tracking continuity even when direct measurements falter. This is crucial for dynamic environments where traditional methods struggle. Our contributions include: 1) Versatile Measurement Augmentation Using KF Feedback: We implement a versatile measurement augmentation system that serves as a backup when primary object detectors fail intermittently. This system is adaptable to various sensors, demonstrated using depth cameras where KF's 3D predictions are projected into 2D depth image coordinates, integrating nonlinear covariance propagation techniques simplified to first-order approximations. 2) Open-source ROS2 Implementation: We provide an open-source ROS2 implementation of the SMART-TRACK framework, validated in a realistic simulation environment using Gazebo and ROS2, fostering broader adaptation and further research. Our results showcase significant enhancements in tracking stability, with estimation RMSE as low as 0.04 m during measurement disruptions, advancing the robustness of UAV tracking and expanding the potential for reliable autonomous UAV operations in complex scenarios. The implementation is available at https://github.com/mzahana/SMART-TRACK., Comment: 12 pages, 7 figures, 3 algorithms, 2 tables

Knee osteoarthritis (KOA) is a prevalent musculoskeletal disorder, often diagnosed using X-rays due to its cost-effectiveness. While Magnetic Resonance Imaging (MRI) provides superior soft tissue visualization and serves as a valuable supplementary diagnostic tool, its high cost and limited accessibility significantly restrict its widespread use. To explore the feasibility of bridging this imaging gap, we conducted a feasibility study leveraging a diffusion-based model that uses an X-ray image as conditional input, alongside target depth and additional patient-specific feature information, to generate corresponding MRI sequences. Our findings demonstrate that the MRI volumes generated by our approach is visually closer to real MRI scans. Moreover, increasing inference steps enhances the continuity and smoothness of the synthesized MRI sequences. Through ablation studies, we further validate that integrating supplementary patient-specific information, beyond what X-rays alone can provide, enhances the accuracy and clinical relevance of the generated MRI, which underscores the potential of leveraging external patient-specific information to improve the MRI generation. This study is available at https://zwang78.github.io/.

Knee Osteoarthritis (KOA) is a common musculoskeletal disorder that significantly affects the mobility of older adults. In the medical domain, images containing temporal data are frequently utilized to study temporal dynamics and statistically monitor disease progression. While deep learning-based generative models for natural images have been widely researched, there are comparatively few methods available for synthesizing temporal knee X-rays. In this work, we introduce a novel deep-learning model designed to synthesize intermediate X-ray images between a specific patient's healthy knee and severe KOA stages. During the testing phase, based on a healthy knee X-ray, the proposed model can produce a continuous and effective sequence of KOA X-ray images with varying degrees of severity. Specifically, we introduce a Diffusion-based Morphing Model by modifying the Denoising Diffusion Probabilistic Model. Our approach integrates diffusion and morphing modules, enabling the model to capture spatial morphing details between source and target knee X-ray images and synthesize intermediate frames along a geodesic path. A hybrid loss consisting of diffusion loss, morphing loss, and supervision loss was employed. We demonstrate that our proposed approach achieves the highest temporal frame synthesis performance, effectively augmenting data for classification models and simulating the progression of KOA.

Biometric recognition is the process of verifying or classifying human characteristics in images or videos. It is a complex task that requires machine learning algorithms, including convolutional neural networks (CNNs) and Siamese networks. Besides, there are several limitations to consider when using these algorithms for image verification and classification tasks. In fact, training may be computationally intensive, requiring specialized hardware and significant computational resources to train and deploy. Moreover, it necessitates a large amount of labeled data, which can be time-consuming and costly to obtain. The main advantage of the proposed TinySiamese compared to the standard Siamese is that it does not require the whole CNN for training. In fact, using a pre-trained CNN as a feature extractor and the TinySiamese to learn the extracted features gave almost the same performance and efficiency as the standard Siamese for biometric verification. In this way, the TinySiamese solves the problems of memory and computational time with a small number of layers which did not exceed 7. It can be run under low-power machines which possess a normal GPU and cannot allocate a large RAM space. Using TinySiamese with only 8 GO of memory, the matching time decreased by 76.78% on the B2F (Biometric images of Fingerprints and Faces), FVC2000, FVC2002 and FVC2004 while the training time for 10 epochs went down by approximately 93.14% on the B2F, FVC2002, THDD-part1 and CASIA-B datasets. The accuracy of the fingerprint, gait (NM-angle 180 degree) and face verification tasks was better than the accuracy of a standard Siamese by 0.87%, 20.24% and 3.85% respectively. TinySiamese achieved comparable accuracy with related works for the fingerprint and gait classification tasks.

Knee OsteoArthritis (KOA) is a widespread musculoskeletal disorder that can severely impact the mobility of older individuals. Insufficient medical data presents a significant obstacle for effectively training models due to the high cost associated with data labelling. Currently, deep learning-based models extensively utilize data augmentation techniques to improve their generalization ability and alleviate overfitting. However, conventional data augmentation techniques are primarily based on the original data and fail to introduce substantial diversity to the dataset. In this paper, we propose a novel approach based on the Vision Transformer (ViT) model with original Selective Shuffled Position Embedding (SSPE) and key-patch exchange strategies to obtain different input sequences as a method of data augmentation for early detection of KOA (KL-0 vs KL-2). More specifically, we fix and shuffle the position embedding of key and non-key patches, respectively. Then, for the target image, we randomly select other candidate images from the training set to exchange their key patches and thus obtain different input sequences. Finally, a hybrid loss function is developed by incorporating multiple loss functions for different types of the sequences. According to the experimental results, the generated data are considered valid as they lead to a notable improvement in the model's classification performance.

Knee Osteoarthritis (KOA) is a prevalent musculoskeletal disorder that severely impacts mobility and quality of life, particularly among older adults. Its diagnosis often relies on subjective assessments using the Kellgren-Lawrence (KL) grading system, leading to variability in clinical evaluations. To address these challenges, we propose a confidence-driven deep learning framework for early KOA detection, focusing on distinguishing KL-0 and KL-2 stages. The Siamese-based framework integrates a novel multi-level feature extraction architecture with a hybrid loss strategy. Specifically, multi-level Global Average Pooling (GAP) layers are employed to extract features from varying network depths, ensuring comprehensive feature representation, while the hybrid loss strategy partitions training samples into high-, medium-, and low-confidence subsets. Tailored loss functions are applied to improve model robustness and effectively handle uncertainty in annotations. Experimental results on the Osteoarthritis Initiative (OAI) dataset demonstrate that the proposed framework achieves competitive accuracy, sensitivity, and specificity, comparable to those of expert radiologists. Cohen's kappa values (k > 0.85)) confirm substantial agreement, while McNemar's test (p > 0.05) indicates no statistically significant differences between the model and radiologists. Additionally, Confidence distribution analysis reveals that the model emulates radiologists' decision-making patterns. These findings highlight the potential of the proposed approach to serve as an auxiliary diagnostic tool, enhancing early KOA detection and reducing clinical workload.

Knee Osteoarthritis (KOA) is a common musculoskeletal condition that significantly affects mobility and quality of life, particularly in elderly populations. However, training deep learning models for early KOA classification is often hampered by the limited availability of annotated medical datasets, owing to the high costs and labour-intensive nature of data labelling. Traditional data augmentation techniques, while useful, rely on simple transformations and fail to introduce sufficient diversity into the dataset. To address these challenges, we propose the Key-Exchange Convolutional Auto-Encoder (KECAE) as an innovative Artificial Intelligence (AI)-based data augmentation strategy for early KOA classification. Our model employs a convolutional autoencoder with a novel key-exchange mechanism that generates synthetic images by selectively exchanging key pathological features between X-ray images, which not only diversifies the dataset but also ensures the clinical validity of the augmented data. A hybrid loss function is introduced to supervise feature learning and reconstruction, integrating multiple components, including reconstruction, supervision, and feature separation losses. Experimental results demonstrate that the KECAE-generated data significantly improve the performance of KOA classification models, with accuracy gains of up to 1.98% across various standard and state-of-the-art architectures. Furthermore, a clinical validation study involving expert radiologists confirms the anatomical plausibility and diagnostic realism of the synthetic outputs. These findings highlight the potential of KECAE as a robust tool for augmenting medical datasets in early KOA detection.

Galaxy clusters are useful laboratories to investigate the evolution of the Universe, and accurately measuring their total masses allows us to constrain important cosmological parameters. However, estimating mass from observations that use different methods and spectral bands introduces various systematic errors. This paper evaluates the use of a Convolutional Neural Network (CNN) to reliably and accurately infer the masses of galaxy clusters from the Compton-y parameter maps provided by the Planck satellite. The CNN is trained with mock images generated from hydrodynamic simulations of galaxy clusters, with Planck's observational limitations taken into account. We observe that the CNN approach is not subject to the usual observational assumptions, and so is not affected by the same biases. By applying the trained CNNs to the real Planck maps, we find cluster masses compatible with Planck measurements within a 15% bias. Finally, we show that this mass bias can be explained by the well known hydrostatic equilibrium assumption in Planck masses, and the different parameters in the Y500-M500 scaling laws. This work highlights that CNNs, supported by hydrodynamic simulations, are a promising and independent tool for estimating cluster masses with high accuracy, which can be extended to other surveys as well as to observations in other bands., Comment: 17 pages, 3 Figures and Supplementary material (+11 figures). Published in Nature Astronomy

We develop a machine learning algorithm to infer the 3D cumulative radial profiles of total and gas mass in galaxy clusters from thermal Sunyaev-Zel'dovich effect maps. We generate around 73,000 mock images along various lines of sight using 2,522 simulated clusters from the \thethreehundred{} project at redshift $z< 0.12$ and train a model that combines an autoencoder and a random forest. Without making any prior assumptions about the hydrostatic equilibrium of the clusters, the model is capable of reconstructing the total mass profile as well as the gas mass profile, which is responsible for the SZ effect. We show that the recovered profiles are unbiased with a scatter of about $10\%$, slightly increasing towards the core and the outskirts of the cluster. We selected clusters in the mass range of $10^{13.5} \leq M_{200} /(\hMsun) \leq 10^{15.5}$, spanning different dynamical states, from relaxed to disturbed halos. We verify that both the accuracy and precision of this method show a slight dependence on the dynamical state, but not on the cluster mass. To further verify the consistency of our model, we fit the inferred total mass profiles with an NFW model and contrast the concentration values with those of the true profiles. We note that the inferred profiles are unbiased for higher concentration values, reproducing a trustworthy mass-concentration relation. The comparison with a widely used mass estimation technique, such as hydrostatic equilibrium, demonstrates that our method recovers the total mass that is not biased by non-thermal motions of the gas., Comment: Accepted for publication on Monthly Notices of the Royal Astronomical Society, 8 pages, 10 figures

Outsourcing anomaly detection to third-parties can allow data owners to overcome resource constraints (e.g., in lightweight IoT devices), facilitate collaborative analysis (e.g., under distributed or multi-party scenarios), and benefit from lower costs and specialized expertise (e.g., of Managed Security Service Providers). Despite such benefits, a data owner may feel reluctant to outsource anomaly detection without sufficient privacy protection. To that end, most existing privacy solutions would face a novel challenge, i.e., preserving privacy usually requires the difference between data entries to be eliminated or reduced, whereas anomaly detection critically depends on that difference. Such a conflict is recently resolved under a local analysis setting with trusted analysts (where no outsourcing is involved) through moving the focus of differential privacy (DP) guarantee from "all" to only "benign" entries. In this paper, we observe that such an approach is not directly applicable to the outsourcing setting, because data owners do not know which entries are "benign" prior to outsourcing, and hence cannot selectively apply DP on data entries. Therefore, we propose a novel iterative solution for the data owner to gradually "disentangle" the anomalous entries from the benign ones such that the third-party analyst can produce accurate anomaly results with sufficient DP guarantee. We design and implement our Differentially Private Outsourcing of Anomaly Detection (DPOAD) framework, and demonstrate its benefits over baseline Laplace and PainFree mechanisms through experiments with real data from different application domains.

The computational study of DNA and its interaction with ligands is a highly relevant area of research, with significant consequences for developing new therapeutic strategies. However, the computational description of such large and complex systems requires considering interactions of different types. All these considerations imply a real challenge for computational chemistry. Using quantum methods for the entire system requires significant computational resources, with improvements in parallelization and optimization of theoretical strategies. Computational methods, such as LS-DFT and DLPNO-CCSD(T), may allow performing ab initio QM calculations, including explicitly the electronic structure for large biological systems, at a reasonable computing time. In this work, we study the interaction of small molecules and cations with DNA (duplex-DNA and G-quadruplexes), comparing different computational methods: a linear-scaling DFT (LS-DFT) at LMKLL/DZDP level of theory, semi-empirical methods (PM6-DH2 and PM7), mixed QM/MM, and DLPNO-CCSD(T). Our goal is to demonstrate the adequacy of LS-DFT to treat the different types of interactions present in DNA-dependent systems. We show that LMKLL/DZDP using SIESTA can yield very accurate geometries and energetics in all the different systems considered in this work: duplex DNA (dDNA), phenanthroline intercalating dDNA, G-quadruplexes, and Metal-G-tetrads considering alkaline metals of different sizes. As far as we know, this is the first time that full G-quadruplex geometry optimizations have been carried out using a DFT method thanks to its linear-scaling capabilities. Moreover, we show that LS-DFT provides high-quality structures, and some semi-empirical Hamiltonian can also yield suitable geometries. However, DLPNO-CCSD(T) and LS-DFT are the only methods that accurately describe interaction energies for all the systems considered in our study., Comment: 32 pages, 6 figures

Clusters of galaxies mass can be inferred by indirect observations, see X-ray band, Sunyaev-Zeldovich (SZ) effect signal or optical. Unfortunately, all of them are affected by some bias. Alternatively, we provide an independent estimation of the cluster masses from the Planck PLSZ2 catalog of galaxy clusters using a machine-learning method. We train a Convolutional Neural Network (CNN) model with the mock SZ observations from The Three Hundred(the300) hydrodynamic simulations to infer the cluster masses from the real maps of the Planck clusters. The advantage of the CNN is that no assumption on a priory symmetry in the cluster's gas distribution or no additional hypothesis about the cluster physical state are made. We compare the cluster masses from the CNN model with those derived by Planck and conclude that the presence of a mass bias is compatible with the simulation results., Comment: To appear in the Proceedings of the International Conference entitled "mm Universe @NIKA2", Rome(Italy), June 2021, EPJ Web of conferences