Search Results (1,007)

Unsupervised person re-identification (Re-ID) has been proven very effective and it boosts the performance in learning representations from unlabeled data in the dataset. Most current methods have good accuracy, but there are two main problems. First, clustering often generates noisy labels. Second, features can change because of different camera styles. Noisy labels causes incorrect optimization, which reduces the accuracy of the model. The latter results in inaccurate prediction for samples within the same category that have been captured by different cameras. Despite the significant variations inherent in the vast source data, the principles of invariance and symmetry remain crucial for effective feature recognition. In this paper, we propose a method called Invariance Constraint Contrast Learning (ICCL) to address these two problems. Specifically, we introduce center invariance and instance invariance to reduce the effect of noisy samples. We also use camera invariance to handle feature changes caused by different cameras. Center invariance and instance invariance help decrease the impact of noise. Camera invariance improves the classification accuracy by using a camera-aware classification strategy. We test our method on three common large-scale Re-ID datasets. It clearly improves the accuracy of unsupervised person Re-ID. Specifically, our approach demonstrates its effectiveness by improving mAP by 3.5% on Market-1501, 1.3% on MSMT17 and 3.5% on CUHK03 over state-of-the-art methods. Full article

Unsupervised medical image anomaly detection, which does not need any labels, holds a pivotal role in early disease detection for advancing human intelligent health, and it is among the prominent research endeavors in the realm of biomedical image analysis. Existing deep model-based methods mainly focus on feature selection and interaction, ignoring the relative position and shape uncertainty of the anomalies themselves, which play an important guiding role in disease diagnosis, hampering performance. To address this issue, our study introduces a novel and effective framework, termed CRTSC, which integrates a channel-wise recalibration module (CRM) along with the texture–structural consistency constraint (TSCC) for anomaly detection in medical chest images acquired from different sensors. Specifically, the CRM adjusts the weight of different medical image feature channels, which are used to establish spatial relationships among anomalous patterns, enhancing the network’s representation and generalization capabilities. The texture–structural consistency constraint is devoted to enhancing the anomaly’s structural (shape) definiteness via evaluating the loss function of similarity between two images and optimizing the model. The two collaborate in an end-to-end fashion to optimize and train the entire framework, thereby enabling anomaly detection in medical chest images. Extensive experiments conducted on the public ZhangLab and CheXpert datasets demonstrate that our method achieves a significant performance improvement compared with the state-of-the-art methods, offering a robust and generalizable solution for sensor-based medical imaging applications. Full article

Geographic Named Entity Recognition (Geo-NER) is a crucial task for extracting geography-related entities from unstructured text, and it plays an essential role in geographic information extraction and spatial semantic understanding. Traditional approaches typically treat Geo-NER as a sequence labeling problem, where each token is assigned a single label. However, this formulation struggles to handle nested entities effectively. To overcome this limitation, we propose Geo-MRC, an improved model based on a Machine Reading Comprehension (MRC) framework that reformulates Geo-NER as a question-answering task. The model identifies entities by predicting their start positions, end positions, and lengths, enabling precise detection of overlapping and nested entities. Specifically, it constructs a unified input sequence by concatenating a type-specific question (e.g., “What are the location names in the text?”) with the context. This sequence is encoded using BERT, followed by feature extraction and fusion through Gated Recurrent Units (GRU) and multi-scale 1D convolutions, which improve the model’s sensitivity to both multi-level semantics and local contextual information. Finally, a feed-forward neural network (FFN) predicts whether each token corresponds to the start or end of an entity and estimates the span length, allowing for dynamic inference of entity boundaries. Experimental results on multiple public datasets demonstrate that Geo-MRC consistently outperforms strong baselines, with particularly significant gains on datasets containing nested entities. Full article

The electrocardiogram (ECG) signal carries information crucial for health assessment, but its analysis can be challenging due to noise and signal variability; therefore, automated processing focused on noise removal and detection of key features is necessary. This paper introduces an ECG signal analysis and abnormality detection application developed to process single-lead ECG signals. In this study, the Lobachevsky University database (LUDB) was used as the source of ECG signals, as it includes annotated recordings using a multi-class, multi-label taxonomy that covers several diagnostic categories, each with specific diagnoses that reflect clinical ECG interpretation practices. The main aim of the paper is to provide a tool that efficiently filters noisy ECG data, accurately detects the QRS complex, PQ and QT intervals, calculates heart rate, and compares these values with normal ranges based on age and gender. Additionally, a multi-class, multi-label SVM-based model was developed and integrated into the application for heart abnormality diagnostics, i.e., assigning one or several diagnoses from various diagnostic categories. The MATLAB-based application is capable of processing raw ECG signals, allowing the use of ECG records not only from LUDB but also from other databases. Full article

Multimodal remote sensing image registration faces severe challenges due to geometric and radiometric differences, particularly between optical and synthetic aperture radar (SAR) images. These inherent disparities make extracting highly repeatable cross-modal feature points difficult. Current methods typically rely on image intensity extreme responses or network regression without keypoint supervision for feature point detection. Moreover, they not only lack explicit keypoint annotations as supervision signals but also fail to establish a clear and consistent definition of what constitutes a reliable feature point in cross-modal scenarios. To overcome this limitation, we propose PLISA—a novel heterogeneous image registration method. PLISA integrates two core components: an automated pseudo-labeling module (APLM) and a pseudo-twin interaction network (PTIF). The APLM introduces an innovative labeling strategy that explicitly defines keypoints as corner points, thereby generating consistent pseudo-labels for dual-modality images and effectively mitigating the instability caused by the absence of supervised keypoint annotations. These pseudo-labels subsequently train the PTIF, which adopts a pseudo-twin architecture incorporating a cross-modal interactive attention (CIA) module to effectively reconcile cross-modal commonalities and distinctive characteristics. Evaluations on the SEN1-2 dataset and OSdataset demonstrate PLISA’s state-of-the-art cross-modal feature point repeatability while maintaining robust registration accuracy across a range of challenging conditions, including rotations, scale variations, and SAR-specific speckle noise. Full article

Event cameras provide microsecond temporal resolution, high dynamic range, and low latency by asynchronously capturing per-pixel luminance changes, thereby introducing a novel sensing paradigm. These advantages render them well-suited for high-speed applications such as autonomous vehicles and dynamic environments. Nevertheless, the sparsity of event data and the absence of dense annotations are significant obstacles to supervised learning for motion segmentation from event streams. Domain adaptation is also challenging due to the considerable domain shift in intensity images. To address these challenges, we propose a two-phase cross-modality adaptation framework that translates motion segmentation knowledge from labeled RGB-flow data to unlabeled event streams. A dual-branch encoder extracts modality-specific motion and appearance features from RGB and optical flow in the source domain. Using reconstruction networks, event voxel grids are converted into pseudo-image and pseudo-flow modalities in the target domain. These modalities are subsequently re-encoded using frozen RGB-trained encoders. Multi-level consistency losses are implemented on features, predictions, and outputs to enforce domain alignment. Our design enables the model to acquire domain-invariant, semantically rich features through the use of shallow architectures, thereby reducing training costs and facilitating real-time inference with a lightweight prediction path. The proposed architecture, alongside the utilized hybrid loss function, effectively bridges the domain and modality gap. We evaluate our method on two challenging benchmarks: EVIMO2, which incorporates real-world dynamics, high-speed motion, illumination variation, and multiple independently moving objects; and MOD++, which features complex object dynamics, collisions, and dense 1kHz supervision in synthetic scenes. The proposed UDA framework achieves 83.1% and 79.4% accuracy on EVIMO2 and MOD++, respectively, outperforming existing state-of-the-art approaches, such as EV-Transfer and SHOT, by up to 3.6%. Additionally, it is lighter and faster and also delivers enhanced mIoU and F1 Score. Full article

Generally, the interpretability of deep neural networks is categorized into a priori and a posteriori interpretability. A priori interpretability involves improving model transparency through deliberate design prior to training. Feature disentanglement is a method for achieving a priori interpretability. Existing disentanglement methods mostly focus on semantic features, such as intrinsic and shared features. These methods distinguish between the background and the main subject, but overlook class-level features in images. To address this, we take a further step by advancing feature disentanglement to the class level. For multi-label image classification tasks, we propose a class-level feature disentanglement method. Specifically, we introduce a multi-head classifier within the feature extraction layer of the backbone network to disentangle features. Each head in this classifier corresponds to a specific class and generates independent predictions, thereby guiding the model to better leverage the intrinsic features of each class and improving multi-label classification precision. Experiments demonstrate that our method significantly enhances performance metrics across various benchmarks while simultaneously achieving a priori interpretability. Full article

Staining-based assays are widely used for cell analysis but are invasive, alter physiology, and prevent longitudinal monitoring. Label-free, morphology-based approaches could enable real-time, non-invasive drug testing, yet detection of subtle and dynamic changes has remained difficult. We developed a deep learning framework for stain-free monitoring of leukemia cell cultures using automated bright-field microscopy in a semi-automated culture system (AICE3, LABMaiTE, Augsburg, Germany). YOLOv8 models were trained on images from K562, HL-60, and Kasumi-1 cells, using an NVIDIA DGX A100 GPU for training and tested on GPU and CPU environments for real-time performance. Comparative benchmarking with RT-DETR and interpretability analyses using Eigen-CAM and radiomics (RedTell) was performed. YOLOv8 achieved high accuracy (mAP@0.5 > 98%, precision/sensitivity > 97%), with reproducibility confirmed on an independent dataset from a second laboratory and an AICE3 setup. The model distinguished between morphologically similar leukemia lines and reliably classified untreated versus differentiated K562 cells (hemin-induced erythroid and PMA-induced megakaryocytic; >95% accuracy). Incorporation of decitabine-treated cells demonstrated applicability to drug testing, revealing treatment-specific and intermediate phenotypes. Longitudinal monitoring captured culture- and time-dependent drift, enabling separation of temporal from drug-induced changes. Radiomics highlighted interpretable features such as size, elongation, and texture, but with lower accuracy than the deep learning approach. To our knowledge, this is the first demonstration that deep learning resolves subtle, drug-induced, and time-dependent morphological changes in unstained leukemia cells in real time. This approach provides a robust, accessible framework for label-free longitudinal drug testing and establishes a foundation for future autonomous, feedback-driven platforms in precision oncology. Ultimately, this approach may also contribute to more precise and adaptive clinical decision-making, advancing the field of personalized medicine. Full article

The segmentation and classification of color are crucial stages in image processing, computer vision, and pattern recognition, as they significantly impact the results. The diverse, hand-labeled datasets in the literature are applied for monochromatic or color segmentation in specific domains. On the other hand, synthetic datasets are generated using statistics, artificial intelligence algorithms, or generative artificial intelligence (AI). This last one includes Large Language Models (LLMs), Generative Adversarial Neural Networks (GANs), and Variational Autoencoders (VAEs), among others. In this work, we propose VitralColor-12, a synthetic dataset for color classification and segmentation, comprising twelve colors: black, blue, brown, cyan, gray, green, orange, pink, purple, red, white, and yellow. VitralColor-12 addresses the limitations of color segmentation and classification datasets by leveraging the capabilities of LLMs, including adaptability, variability, copyright-free content, and lower-cost data—properties that are desirable in image datasets. VitralColor-12 includes pixel-level classification and segmentation maps. This makes the dataset broadly applicable and highly variable for a range of computer vision applications. VitralColor-12 utilizes GPT-5 and DALL·E 3 for generating stained-glass images. These images simplify the annotation process, since stained-glass images have isolated colors with distinct boundaries within the steel structure, which provide easy regions to label with a single color per region. Once we obtain the images, we use at least one hand-labeled centroid per color to automatically cluster all pixels based on Euclidean distance and morphological operations, including erosion and dilation. This process enables us to automatically label a classification dataset and generate segmentation maps. Our dataset comprises 910 images, organized into 70 generated images and 12 pixel segmentation maps—one for each color—which include 9,509,524 labeled pixels, 1,794,758 of which are unique. These annotated pixels are represented by RGB, HSL, CIELAB, and YCbCr values, enabling a detailed color analysis. Moreover, VitralColor-12 offers features that address gaps in public resources such as violin diagrams with the frequency of colors across images, histograms of channels per color, 3D color maps, descriptive statistics, and standardized metrics, such as ΔE76, ΔE94, and CIELAB Chromacity, which prove the distribution, applicability, and realistic perceptual structures, including warm, neutral, and cold colors, as well as the high contrast between black and white colors, offering meaningful perceptual clusters, reinforcing its utility for color segmentation and classification. Full article

Although current 3D semantic segmentation methods have achieved significant success, they suffer from catastrophic forgetting when confronted with dynamic, open environments. To address this issue, class incremental learning is introduced to update models while maintaining a balance between plasticity and stability. In this work, we propose CosPrompt, a rehearsal-free approach for class incremental semantic segmentation. Specifically, we freeze the prompts for existing classes and incrementally expand and fine-tune the prompts for new classes, thereby generating discriminative and customized features. We employ clamping operations to regulate backward propagation, ensuring smooth training. Furthermore, we utilize the learning without forgetting loss and pseudo-label generation to further mitigate catastrophic forgetting. We conduct comparative and ablation experiments on the S3DIS dataset and ScanNet v2 dataset, demonstrating the effectiveness and feasibility of our method. Full article

Agricultural management in Europe faces increasing pressure to reduce its environmental footprint. Implementing precision agriculture for weed management could offer a solution and minimize the use of chemical products. High spatial resolution imagery from real time kinematic (RTK) unmanned aerial vehicles (UAV) in combination with supervised convolutional neural network (CNNs) models have proven successful in making location specific treatments. This site-specific advice limits the amount of herbicide applied to the field to areas that require action, thereby reducing the environmental impact and inputs for the farmer. To develop performant CNN models, there is a need for sufficient high-quality labelled data. To reduce the labelling effort and time, a new labelling method is proposed whereby image subsection pairs are labelled based on their relative differences in weed pressure to train a CNN ordinal regression model. The model is evaluated on detecting weed pressure in potato (Solanum tuberosum L.). Model performance was evaluated on different levels: pairwise accuracy, linearity (Pearson correlation coefficient), rank consistency (Spearman’s (r_s) and Kendal (τ) rank correlations coefficients) and binary accuracy. After hyperparameter tuning, a pairwise accuracy of 85.2%, significant linearity (r_s = 0.81) and significant rank consistency (r_s = 0.87 and τ = 0.69) were found. This suggests that the model is capable of correctly detecting the gradient in weed pressure for the dataset. A maximum binary accuracy and F1-score of 92% and 88% were found for the dataset after thresholding the predicted weed scores into weed versus non-weed images. The model architecture allows us to visualize the intermediate features of the last convolutional block. This allows data analysts to better evaluate if the model “sees” the features of interest (in this case weeds). The results indicate the potential of ordinal regression with relative labels as a fast, lightweight model that predicts weed pressure gradients. Experts have the freedom to decide which threshold value(s) can be used on predicted weed scores depending on the weed, crop and treatment that they want to use for flexible weed control management. Full article

Accurate and detailed tissue characterization is a central goal in medical diagnostics, often requiring the combination of multiple imaging modalities. This study presents a multimodal imaging system that integrates mid-infrared (MIR) scanning with fluorescence imaging to enhance the chemical specificity and spatial resolution in biological samples. A motorized mirror allows rapid switching between MIR and fluorescence modes, enabling efficient, co-registered data acquisition. The MIR modality captures label-free chemical maps based on molecular vibrations, while the fluorescence channel records endogenous autofluorescence for additional biochemical contrast. Applied to mouse brain tissue, the system enabled the clear differentiation of gray matter and white matter, supported by the clustering analysis of spectral features. The addition of autofluorescence imaging further improved anatomical segmentation and revealed fine structural details. In mouse skin, the approach allowed the precise mapping of the layered tissue architecture. These results demonstrate that combining MIR scanning and fluorescence imaging provides complementary, label-free insights into tissue morphology and chemistry. The findings support the utility of this approach as a powerful tool for biomedical research and diagnostic applications, offering a more comprehensive understanding of tissue composition without relying on staining or external markers. Full article

The classification of urban residential settlements through Machine Learning (ML) and Deep Learning (DL) remains a complex task due to the intrinsic heterogeneity of urban environments and the scarcity of large, accurately labeled training datasets. To overcome these limitations, this study introduces a novel GIS-based unsupervised classification framework that exploits Fuzzy C-Means (FCM) clustering for the detection and interpretation of urban morphologies. Compared to unsupervised classification approaches that rely on crisp-based clustering algorithms, the proposed FCM-based method more effectively captures heterogeneous urban fabrics where no clear predominance of specific building types exists. Specifically, the method applies fuzzy clustering to census units—considered the fundamental scale of urban analysis—based on construction techniques and building periods. By grouping census areas with similar structural features, the framework provides a flexible, data-driven approach to the characterization of urban settlements. The identification of cluster centroids’ dominant attributes enables a systematic interpretation of the spatial distribution of the built environment, while the subsequent mapping process assigns each cluster a descriptive label reflecting the prevailing building fabric. The generated thematic maps yield critical insights into urban morphology and facilitate evidence-based planning. The framework was validated across ten Italian cities selected for their diverse physical, morphological, and historical characteristics; comparisons with the results of urban zone classifications in these cities conducted by experts show that the proposed method provides accurate results, as the similarity to the classifications made by experts, measured by the use of the Adjusted Rand Index, is always higher than or equal to 0.93; furthermore, it is robust when applied in heterogeneous urban settlements. These results confirm the effectiveness of the method in delineating homogeneous urban areas, thereby offering decision makers a robust instrument to guide targeted interventions on existing building stocks. The proposed framework advances the capacity to analyze urban form, to strategically support renovation and urban regeneration policies, and demonstrates a strong potential for portability, as it can be applied to other cities for urban scale analyses. Full article

Accurate bearing fault diagnosis under various operational conditions presents significant challenges, mainly due to the limited availability of labeled data and the domain mismatches across different operating environments. In this study, an adaptive meta-learning framework (AdaMETA) is proposed, which combines dynamic task-aware model-independent meta-learning (DT-MAML) with efficient multi-scale attention (EMA) modules to enhance the model’s ability to generalize and improve diagnostic performance in small-sample bearing fault diagnosis across different load scenarios. Specifically, a hierarchical encoder equipped with C-EMA is introduced to effectively capture multi-scale fault features from vibration signals, greatly improving feature extraction under constrained data conditions. Furthermore, DT-MAML dynamically adjusts the inner-loop learning rate based on task complexity, promoting efficient adaptation to diverse tasks and mitigating domain bias. Comprehensive experimental evaluations on the CWRU bearing dataset, conducted under carefully designed cross-domain scenarios, demonstrate that AdaMETA achieves superior accuracy (up to 99.26%) and robustness compared to traditional meta-learning and classical diagnostic methods. Additional ablation studies and noise interference experiments further validate the substantial contribution of the EMA module and the dynamic learning rate components. Full article

Achieving a dataset of eye detection comprises a critical task in computer vision and image processing. The primary goal of this dataset is to accurately locate and identify the position of eyes in image or video frames. This process can firstly detect the face region and then focus on the eye regions. In this study, 14,980 examples of physiological signal recordings, most likely from EEG or similar sensors, were included in this dataset, which was created for the analysis of neural or sensor-based movement. The constant signals from specific sensor channels are represented by 14 numerical features (AF3, F7, F3, O1, O2, P7, P8, T8, FC5, FC6, etc.). These characteristics record complex changes in signal designs over time, which could suggest shifts in sensor or neuronal activity. Also, the dataset involves a binary target variable called eye detection, and this shows if an eye-related event—such as turning or an open/closed state—is identified during an individual case. The basic label of this dataset is eye detection in human beings, which has instances of (0,1). The eye detection dataset has 14 features and 14,980 instances that can be utilized for training a model. Full article