Sparse local feature matching has served as the cornerstone of numerous visual geometry tasks and attracted extensive attention. Although significant progress has been made in this area, improving the discriminative power of descriptors remains a key challenge. As far as we know, existing sparse feature matching methods only predict a single descriptor map for keypoints, which might restrict their potential in solving complex scenarios. This issue is particularly pronounced in real-time applications where most methods only learn descriptor maps at a reduced spatial resolution compared to the input image. Consequently, they require interpolating from the low resolution map for obtaining per-keypoint descriptors, which will introduce background contamination and reduce the discriminability of final descriptors. To address these issues, we propose an efficient novel complementary local feature description model. Specifically, the model simultaneously learns two descriptor maps using different loss functions within a single Convolutional Neural Network (CNN). An orthogonal loss is introduced to effectively coordinate the learning of the two branches, aiming to obtain decoupled and complementary descriptors. Extensive experiments across various visual geometry tasks, such as homography estimation, indoor and outdoor pose estimation, as well as visual localization, have demonstrated the superior performance of the proposed method.

Artificial intelligence (AI) is increasingly applied in pulmonary endoscopy, including diagnostic bronchoscopy, interventional pulmonology and endobronchial imaging. Advances in computer vision, machine learning and robotic systems have expanded the potential for automated lesion detection, navigation to peripheral pulmonary lesions, and real-time procedural support. However, the current evidence base remains heterogeneous, and translational challenges persist. This review summarizes current applications and developments of AI across white-light bronchoscopy (WLB), image-enhanced bronchoscopy (e.g., narrow-band imaging and autofluorescence imaging), endobronchial ultrasound (EBUS), virtual and robotic bronchoscopies, and workflow optimization and training. The authors also examine the methodological limitations, regulatory considerations, and implementation barriers that affect translation into routine practice. Reported developments include deep learning-based models for mucosal abnormality detection, lymph-node characterization during EBUS-guided transbronchial needle aspiration (EBUS-TBNA), improved lesion localization, and reduction in operator-dependent variability. Additionally, AI-assisted simulation platforms and decision-support tools are reshaping training paradigms. Nevertheless, most studies remain retrospective or single-center, with limited external validation, dataset heterogeneity, unclear model explainability, and incomplete integration into clinical workflows. AI has the potential to support lesion detection, navigation, and training in pulmonary endoscopy. However, robust prospective validation, standardized datasets, transparent model reporting, robust data governance, multidisciplinary collaboration, and careful integration into clinical practice are required before widespread adoption.

Ultrasound images have some limitations, such as low signal-to-noise ratio (SNR), speckle noise, lower dynamic range, blurred boundaries, and shadowing; therefore, ultrasound image registration is an important task for estimating tissue motion and analyzing tissue mechanical properties. In this paper, an effective non-rigid ultrasound image registration method is proposed. By integrating intensity, local phase information, and descriptor matching under a variational framework, we can find and track the non-rigid transformation of each pixel under diffeomorphism between the source and target images based on the warping technique. Experiments using simulation and in vivo ultrasound images of the human carotid artery are conducted to demonstrate the advantages of the proposed algorithm, which will act as an important supplement to current ultrasound image registration.

Segment Anything Model (SAM)-based few-shot segmentation models traditionally rely solely on annotated reference images as prompts, which inherently limits their accuracy due to an over-reliance on visual cues and a lack of semantic context. This reliance leads to incorrect segmentation, where visually similar objects from different categories are incorrectly identified as the target object. We propose Vision and Language Reference Prompt into SAM (VLP-SAM), a novel few-shot segmentation model that integrates both visual information of reference images and semantic information of text labels into SAM. VLP-SAM introduces a vision-language model (VLM) with pixel-text matching into the prompt encoder for SAM, effectively leveraging textual semantic consistency while preserving SAM's extensive segmentation knowledge. By incorporating task-specific structures such as an attention mask, our model achieves superior few-shot segmentation performance with only 1.4 M learnable parameters. Evaluations on PASCAL-5i and COCO-20i datasets demonstrate that VLP-SAM significantly outperforms previous methods by 6.8% and 9.3% in mIoU, respectively. Furthermore, VLP-SAM exhibits strong generalization across unseen objects and cross-domain scenarios, highlighting the robustness provided by textual semantic guidance. This study offers an effective and scalable framework for few-shot segmentation with multimodal prompts.

Synthetic dataset generation in Computer Vision, particularly for industrial applications, is still underexplored. Industrial defect segmentation, for instance, requires highly accurate labels, yet acquiring such data is costly and time-consuming. To address this challenge, we propose a novel diffusion-based pipeline for generating high-fidelity industrial datasets with minimal supervision. Our approach conditions the diffusion model on enriched bounding-box representations to produce precise segmentation masks, ensuring realistic and accurately localized defect synthesis. Compared to existing layout-conditioned generative methods, our approach improves defect consistency and spatial accuracy. We introduce two quantitative metrics to evaluate the effectiveness of our method and assess its impact on a downstream segmentation task trained on real and synthetic data. Our results demonstrate that diffusion-based synthesis can bridge the gap between artificial and real-world industrial data, fostering more reliable and cost-efficient segmentation models.

This study assessed the level of agreement between quantitative ultrasound (QUS) feature estimates derived from ultrasound images of breast tumors in women with locally advanced breast cancer (LABC) produced using a cart-based and a handheld ultrasound system. Thirty LABC patients receiving neoadjuvant chemotherapy were imaged at two separate times: a pre-treatment 'baseline' time point, and four weeks after the start of chemotherapy. Three sets of QUS features were produced using the reference phantom technique, one for each imaging time and a third set calculated by taking the differences in feature estimates between times. Cross-system statistical testing using the Wilcoxon signed-rank test was performed for each feature set to assess the level of feature estimate agreement between ultrasound systems. The Bland-Altman method was employed to graphically assess feature sets for systematic skew. The range of p-values was 4.50 × 10-11 to 0.277 for the baseline features, 2.77 × 10-5 to 0.865 for the week 4 features, and 2.03 × 10-9 to 1 for the feature differences. For the feature differences, all five of the primary QUS features (MBF, SS, SI, ASD, AAC) were found to be in agreement between the two scanner types at the 5% confidence level. For the baseline feature set and week 4 feature set, 0 out of 5 and 3 out of 5 of the primary features were found to be in agreement, respectively. Of the 20 QUS texture features examined, the number and proportion of the total for each feature set which were found to have statistically significant similarity in their sample medians at the 5% confidence level were as follows: 2 out of 20 (10%) for the baseline features; 17 out of 20 (85%) for the week 4 features; and 12 out of 20 (60%) for the feature differences. The specific texture features found to be in agreement varied between QUS-specific feature sets. Overall, a moderate level of agreement between sets of feature differences produced using the two systems was demonstrated.

Accurate color property measurements are critical for advancing artificial vision in real-time industrial applications. RGB imaging remains highly applicable and widely used due to its practicality, accessibility, and high spatial resolution. However, significant uncertainties in extracting chromatic information highlight the need to define when conventional digital images can reliably provide accurate color data. This work simultaneously compares six chromatic properties across 700 Pantone® TCX fabric samples, using optical data acquired simultaneously from both hyperspectral (HSI) and digital (RGB) cameras. The results indicate that the accurate interpretation of optical information from RGB (sRGB and REC2020) images is significantly influenced by lightness (L*) values. Samples with bright and unsaturated colors (L*> 50) reach ratio-to-performance-deviation (RPD) values above 2.5 for four properties (L*, a*, b* hab), indicating a good correlation between HSI and RGB information. Absolute color difference comparisons (∆Ea) between HSI and RGB images yield values exceeding 5.5 units for red-yellow-green samples and up to 9.0 units for blue and purple tones. In contrast, relative color differences (∆Er) comparisons show a significant decrease, with values falling below 3.0 for all lightness values, indicating the practical equivalence of both methodologies according to the Two One-Sided Test (TOST) statistical analysis. These results confirm that RGB imagery achieves reliable color consistency when evaluated against a practical reference.

Head and neck cancer (HNC) management requires highly coordinated multidisciplinary care. Radiation Therapy Technologists (RTTs) have increasingly expanded their role beyond technical execution, contributing to patient positioning, treatment delivery, monitoring, and supportive care. This narrative review integrates evidence from published literature with structured experiential insights collected through focus group discussions with RTTs and other multidisciplinary team (MDT) members. The resulting conceptual and operational framework highlights RTT contributions across the radiotherapy pathway, including adaptive planning, workflow coordination, and patient-centered interventions, supported by imaging and artificial intelligence (AI) tools for predictive modeling and treatment optimization. By facilitating communication, monitoring anatomical and functional changes, and integrating AI-informed insights, RTTs support timely interventions, reduce treatment interruptions, and enhance treatment safety and precision. Structured training, formal recognition of advanced practice roles, and interprofessional collaboration are key to maximizing the impact of RTTs in HNC care. This review provides a practical roadmap for institutions, professional societies, and research initiatives to support the evolution of RTT roles in complex radiotherapy settings.

Source camera identification on smartphones constitutes a fundamental task in multimedia forensics, providing essential support for applications such as image copyright protection, illegal content tracking, and digital evidence verification. Numerous techniques have been developed for this task over the past decades. Among existing approaches, Photo-Response Non-Uniformity (PRNU) has been widely recognized as a reliable device-specific fingerprint and has demonstrated remarkable performance in real-world applications. Nevertheless, the rapid advancement of computational photography technologies has introduced significant challenges: modern devices often exhibit anomalous behaviors under PRNU-based analysis. For instance, images captured by different devices may exhibit unexpected correlations, while images captured by the same device can vary substantially in their PRNU patterns. Current approaches are incapable of automatically exploring the underlying causes of these anomalous behaviors. To address this limitation, we propose a simple yet effective forensic analysis framework leveraging Exchangeable Image File Format (EXIF) metadata. Specifically, we represent EXIF metadata as type-aware word embeddings to preserve contextual information across tags. This design enables visual interpretation of the model's decision-making process and provides complementary insights for identifying the anomalous behaviors observed in modern devices. Extensive experiments conducted on three public benchmark datasets demonstrate that the proposed method not only achieves state-of-the-art performance for source camera identification but also provides valuable insights into anomalous device behaviors.

When capturing star images in long-exposure mode, due to the relative motion between stars and space objects and the observation camera, strip tailings with different directions and lengths will be formed, resulting in a serious decline in image quality and inaccurate centroid positioning. Traditional methods for restoring star images are prone to ringing effects and cannot restore the non-uniformly blurred star images. Aiming at this problem, this paper proposes a star image restoration network based on a dynamic strip attention mechanism. Firstly, a Multi-scale Dynamic Strip Pooling Module is designed to adaptively extract blurred features of different lengths and directions by dynamically adjusting the strip convolution. After that, a Multi-scale Feature Fusion Module is designed to fuse multi-level features to reduce the loss of image details of stars and space objects in the image. Experimental results demonstrate that the proposed method achieves a PSNR of 84.08 and an SSIM of 0.9928 on the 16-bit simulated dataset, outperforming both traditional methods and other deep learning-based approaches. Specifically, the recognition accuracy of star points is increased by 174% in comparison with unprocessed images. Furthermore, this paper validates the network using the real-world dataset spotGEO, and the results indicate that the average number of successfully recognized star points is increased by 57% compared to direct processing of the original images.