Local Image Research Articles

Enzyme-Responsive Fluorescent Labeling Strategy for In Vivo Imaging of Gut Bacteria.

Myrosinase (Myr), as a unique β-thioglucosidase enzyme capable of converting natural and gut bacterial metabolite glucosinolates into bioactive agents, has recently attracted a great deal of attention because of its essential functions in exerting homeostasis dynamics and promoting human health. Such nutraceutical and biomedical significance demands unique and reliable strategies for specific identification of Myr enzymes of gut bacterial origin in living systems, whereas the dearth of methods for bacterial Myr detection and visualization remains a challenging concern. Herein, we present a series of unique molecular probes for specific identification and imaging of Myr-expressing gut bacterial strains. Typically, an artificial glucosinolate with an azide group in aglycone was synthesized and sequentially linked with the probe moieties of versatile channels through simple click conjugation. Upon gut bacterial enzymatic cleavage, the as-prepared probe molecules could be converted into reactive isothiocyanate forms, which can further act as reactive electrophiles for the covalent labeling of gut bacteria, thus realizing their localized fluorescent imaging within a wide range of wavelength channels in live bacterial strains and animal models. Overall, our proposed method presents a novel technology for selective gut bacterial Myr enzyme labeling in vitro and in vivo. We envision that such a rational probe design would serve as a promising solution for chemoprevention assessment, microflora metabolic mechanistic study, and gut bacterium-mediated physiopathological exploration.

Remote Monitoring and Data Collection for Decentralized Clinical Trials

Less than 5% of patients with cancer enroll in a clinical trial, partly due to financial and logistic burdens, especially among underserved populations. The COVID-19 pandemic marked a substantial shift in the adoption of decentralized trial operations by pharmaceutical companies. To assess the current global state of adoption of decentralized trial technologies, understand factors that may be driving or preventing adoption, and highlight aspirations and direction for industry to enable more patient-centric trials. The Bloomberg New Economy International Cancer Coalition, composed of patient advocacy, industry, government regulator, and academic medical center representatives, developed a survey directed to global biopharmaceutical companies of the coalition from October 1 through December 31, 2022, with a focus on registrational clinical trials. The data for this survey study were analyzed between January 1 and 31, 2023. Adoption of decentralized clinical trial technologies. The survey measured (1) outcomes of different remote monitoring and data collection technologies on patient centricity, (2) adoption of these technologies in oncology and all therapeutic areas, and (3) barriers and facilitators to adoption using descriptive statistics. All 8 invited coalition companies completed the survey, representing 33% of the oncology market by revenues in 2021. Across nearly all technologies, adoption in oncology trials lags that of all trials. In the current state, electronic diaries and electronic clinical outcome assessments are the most used technology, with a mean (SD) of 56% (19%) and 51% (29%) adoption for all trials and oncology trials, respectively, whereas visits within local physician networks is the least adopted at a mean (SD) of 12% (18%) and 7% (9%), respectively. Looking forward, the difference between the current and aspired adoption rate in 5 years for oncology is large, with respondents expecting a 40% or greater absolute adoption increase in 8 of the 11 technologies surveyed. Furthermore, digitally enabled recruitment, local imaging capabilities, and local physician networks were identified as technologies that could be most effective for improving patient centricity in the long term. These findings may help to galvanize momentum toward greater adoption of enabling technologies to support a new paradigm of trials that are more accessible, less burdensome, and more inclusive.

Spatio-temporal evolution of sports tourism industry integration: A empirical analysis of Xinjiang in China.

Since the issuance of the "Guiding Opinions on Vigorously Developing Sports Tourism" in 2016, the integration of sports and tourism has become a strategy in regional economic development. It creates new economic growth points, enhances local images, and promotes cultural communication. In the context of the "Tourism Makes Xinjiang Thrive" strategy, quantitatively investigating the integration of the sports and tourism industries helps people to better understand their interaction which can serve as the valuable input in policy-making for the comprehensive development of a region. This paper uses entropy weight method, stochastic frontier analysis and coupling coordination model to quantitatively analyze the effect of sports tourism industry integration in Xinjiang from the perspective of integration path. Meanwhile, the Dagum Gini coefficient and nuclear density estimation were used to analyze the regional differences and dynamic evolution of industrial integration quality. The result shows that (1) The sports and tourism integration quality in Xinjiang has not reached the optimal goal of complete integration. In the process of mutual industrial promotion, tourism promotes a higher degree of integration with the sports industry. (2) The industrial integration quality shows a phenomenon of "imbalance and inadequacy" among the regions. The regions with high quality of industrial integration were Urumqi, Ili, Kashgar, Altay and Changji, which have rich sports tourism resources. (3) The overall spatial difference in the quality of industrial integration presented a fluctuation downtrend. The difference between the tourism industrial belts was very significant.

Non-local degradation modeling for spatially adaptive single image super-resolution

Existing methods for single image super-resolution (SISR) model the blur kernel as spatially invariant across the entire image, and are susceptible to the adverse effects of textureless patches. To achieve improved results, adaptive estimation of the degradation kernel is necessary. We explore the synergy of joint global and local degradation modeling for spatially adaptive blind SISR. Our model, named spatially adaptive network for blind super-resolution (SASR), employs a simple encoder to estimate global degradation representations and a decoder to extract local degradation. These two representations are fused with a cross-attention mechanism and applied using spatially adaptive filtering to enhance the local image detail. Specifically, SASR contains two novel features: (1) a non-local degradation modeling with contrastive learning to learn global and local degradation representations, and (2) a non-local spatially adaptive filtering module (SAFM) that incorporates the global degradation and spatial-detail factors to preserve and enhance local details. We demonstrate that SASR can efficiently estimate degradation representations and handle multiple types of degradation. The local representations avoid the detrimental effect of estimating the entire super-resolved image with only one kernel through locally adaptive adjustments. Extensive experiments are performed to quantitatively and qualitatively demonstrate that SASR not only performs favorably for degradation estimation but also leads to state-of-the-art blind SISR performance when compared to alternative approaches.

Feature Engineering Versus Deep Learning for Scene Recognition: A Brief Survey

Scene recognition is an important computer vision task that has evolved from the study of the biological visual system. Its applications range from video surveillance, autopilot systems, to robotics. The early works were based on feature engineering that involved the computation and aggregation of global and local image descriptors. Several popular image features such as SIFT, SURF, HOG, ORB, LBP, KAZE, etc. have been proposed and applied to the task with successful results. Features can be either computed from the entire image on a global scale, or extracted from local sub-regions and aggregated across the image. Suitable classifier models are deployed that learn to classify these features. This review paper analyzes several of these handcrafted features that have been applied to the scene recognition task over the past decades, and tracks the transition from the traditional feature engineering to deep learning which forms the current state of the art in computer vision. Deep learning is now deemed to have overtaken feature engineering in several computer vision applications. Deep convolutional neural networks and vision transformers are the current state of the art for object recognition. However, scenes from urban landscapes are bound to contain similar objects posing a challenge to deep learning solutions for scene recognition. In our study, a critical analysis of feature engineering and deep learning methodologies for scene recognition is provided, and results on benchmark scene datasets are presented, concluding with a discussion on challenges and possible solutions that may facilitate more accurate scene recognition.

A Distortion Correction Method Based on Actual Camera Imaging Principles.

In the human-robot collaboration system, the high-precision distortion correction of the camera as an important sensor is a crucial prerequisite for accomplishing the task. The traditional correction process is to calculate the lens distortion with the camera model parameters or separately from the camera model. However, in the optimization process calculate with the camera model parameters, the mutual compensation between the parameters may lead to numerical instability, and the existing distortion correction methods separated from the camera model are difficult to ensure the accuracy of the correction. To address this problem, this study proposes a model-independent lens distortion correction method based on the image center area from the perspective of the actual camera lens distortion principle. The proposed method is based on the idea that the structured image preserves its ratios through perspective transformation, and uses the local image information in the central area of the image to correct the overall image. The experiments are verified from two cases of low distortion and high distortion under simulation and actual experiments. The experimental results show that the accuracy and stability of this method are better than other methods in training and testing results.

Condylar fracture location is correlated to exercise history in Thoroughbred racehorses.

Condylar fractures are a major cause of morbidity and mortality in Thoroughbred racehorses. Condylar fractures have a variety of fracture configurations that suggest there may be differences in aetiopathogenesis. To determine if exercise history differs with condylar fracture location in a population of Thoroughbred racehorses. Retrospective analysis of clinical and exercise data. Exercise history of Thoroughbred racehorses that had condylar fracture repair between 1 January 2018 and 28 February 2021 was compared between racehorses that had fractures located radiographically either within the parasagittal groove (PSG) or abaxial to the PSG (non-PSG). Age, sex, and last event (race, timed work) matched control groups were compared between the PSG and non-PSG groups. Additionally, exercise history variables of both groups were each compared with a group-specific control population, each consisting of three control racehorses of equivalent age and sex matched to each affected racehorse by last event (race or official timed work) before fracture. Eighty-two horses with 84 fractures (45 PSG, 39 non-PSG) met inclusion criteria. Age was not different between groups (PSG: 3.4 ± 1.3 years [mean ± SD], non-PSG: 3.7 ± 1.3, p = 0.3). Number of races (PSG: 5.3 ± 7.1, non-PSG: 11.4 ± 8.9, p < 0.001), total race furlongs (PSG: 38.2 ± 54.7, non-PSG: 79.2 ± 64, p = 0.003), and number of active days (PSG: 304 ± 224, non-PSG: 488 ± 314, p = 0.003) before fracture were greater; while mean number of layups was fewer (PSG: 1.0 ± 1.2, non-PSG: 0.5 ± 0.7, p = 0.02) in horses with non-PSG fracture. Horses with non-PSG fracture had more differences compared with their respective control group than horses with PSG fractures. Outcomes following fracture repair were not different between groups. Retrospective study, one regional racehorse population, two-dimensional imaging and potential inherent bias for fracture localisation, low statistical power for return to performance analysis. Thoroughbred racehorses with non-PSG condylar fractures have a more extensive exercise history than horses with PSG condylar fractures, suggesting differences in fracture aetiopathogenesis.

Research on Efficient Feature Generation and Spatial Aggregation for Remote Sensing Semantic Segmentation

Semantic segmentation algorithms leveraging deep convolutional neural networks often encounter challenges due to their extensive parameters, high computational complexity, and slow execution. To address these issues, we introduce a semantic segmentation network model emphasizing the rapid generation of redundant features and multi-level spatial aggregation. This model applies cost-efficient linear transformations instead of standard convolution operations during feature map generation, effectively managing memory usage and reducing computational complexity. To enhance the feature maps’ representation ability post-linear transformation, a specifically designed dual-attention mechanism is implemented, enhancing the model’s capacity for semantic understanding of both local and global image information. Moreover, the model integrates sparse self-attention with multi-scale contextual strategies, effectively combining features across different scales and spatial extents. This approach optimizes computational efficiency and retains crucial information, enabling precise and quick image segmentation. To assess the model’s segmentation performance, we conducted experiments in Changge City, Henan Province, using datasets such as LoveDA, PASCAL VOC, LandCoverNet, and DroneDeploy. These experiments demonstrated the model’s outstanding performance on public remote sensing datasets, significantly reducing the parameter count and computational complexity while maintaining high accuracy in segmentation tasks. This advancement offers substantial technical benefits for applications in agriculture and forestry, including land cover classification and crop health monitoring, thereby underscoring the model’s potential to support these critical sectors effectively.

Machine learning-based automated scan prescription of lumbar spine MRI acquisitions

PurposeHigh quality scan prescription that optimally covers the area of interest with scan planes aligned to relevant anatomical structures is crucial for error-free radiologic interpretation. The goal of this project was to develop a machine learning pipeline for oblique scan prescription that could be trained on localizer images and metadata from previously acquired MR exams. MethodsA novel Multislice Rotational Region-based Convolutional Neural Network (MS-R2CNN) architecture was developed. Based on this architecture, models for automated prescription sagittal lumbar spine acquisitions from axial, sagittal, and coronal localizer slices were trained. The automated prescription pipeline was integrated with the scanner console software and evaluated in experiments with healthy volunteers (N = 3) and patients with lower-back pain (N = 20). ResultsExperiments in healthy volunteers demonstrated high accuracy of automated prescription in all subjects. There was good agreement between alignment and coverage of manual and automated prescriptions, as well as consistent views of the lumbar spine at different positions of the subjects within the scanner bore. In patients with lower-back pain, the generated prescription was applied in 18 cases (90% of the total number). None of the cases required major adjustment, while in 11 cases (55%) there were minor manual adjustments to the generated prescription. ConclusionsThis study demonstrates the ability of oriented object detection-based models to be trained to prescribe oblique lumbar spine MRI acquisitions without the need of manual annotation or feature engineering and the feasibility of using machine learning-based pipelines on the scanner for automated prescription of MRI acquisitions.

Local Adaptive Image Equalization

This paper presents a comprehensive approach to image enhancement, targeting the enhancement of contrast and reduction of noise in digital images. Leveraging state-of-the-art algorithms, the proposed methodology encompasses a strategic pipeline. Initially, the images undergo Histogram Equalization, a fundamental technique, to globally enhance contrast. Building upon this foundation, Contrast Limited Adaptive Histogram Equalization (CLAHE) is applied to achieve localized contrast enhancement, ensuring optimal balance and preservation of image details. Furthermore, the Adaptive Gamma Correction with Weighting Distribution (AGCWD) algorithm is integrated to fine-tune the enhanced images, dynamically adjusting gamma values to suppress noise and amplify visual features. The implementation harnesses Python with OpenCV and Flask frameworks, facilitating seamless integration and accessibility. Through rigorous experimentation and comparative analysis, the efficacy of the proposed approach is demonstrated, showcasing substantial improvements in image quality and fidelity. The findings underscore the practical utility and efficacy of the proposed image enhancement system, positioning it as a valuable tool for various real-world applications in image processing and computer vision domains. Keywords: Image Enhancement ,Noise Reduction ,Histogram Equalization, CLAHE, Adaptive Gamma Correction, OpenCV Library, Flask Web Framework , Image Fidelity.

Reducing noise in polarization-sensitive optical coherence tomography for high-quality local phase retardation imaging

Local phase retardation (LPR) is increasingly recognized as a crucial biomarker for assessing disease progression. However, the presence of speckle noise significantly challenges its accuracy and polarization contrast. To address this challenge, we propose a signal-processing strategy aimed at reducing the impact of noise on LPR measurements. In this approach, the LPR is reconstructed by polar decomposition after averaging multiple Mueller matrices from different overlapping sub-spectra. To optimize measurement accuracy, we systematically combined and traversed different sub-spectral numbers and bandwidths. By examining the quarter-wave plate and glass slide, high-accuracy phase retardation measurements were successfully verified, and the maximum polarization contrast was improved by 23%. Moreover, experimental results from multi-tissue imaging vividly illustrate that the equivalent number of looks (ENL) and polarization contrast were improved by 18% and 19%, respectively. This outcome indicates that our proposed strategy can effectively reduce the noise spikes, enhancing tissue discrimination capabilities.

Alignment and fusion for adaptive domain nighttime semantic segmentation

In the field of autonomous driving technology, both daytime and nighttime scenes are common. However, due to the poor illumination and difficulty in manual annotation of nighttime images, semantic segmentation of nighttime scenes is more challenging compared to daytime scenes. Therefore, achieving significant progress in nighttime semantic segmentation would greatly enhance the effectiveness of the application of autonomous driving scenarios. In this work, we investigate the problem of domain-adaptive nighttime semantic segmentation (DANSS). The problem aims to learn semantic segmentation in nighttime scenes by leveraging a labeled Cityscapes dataset and unlabeled but roughly aligned day-night image pairs. To address this, we propose an Align and Fusion Network (AAFnet), a network for adaptive domain nighttime semantic segmentation. AAFnet utilizes a novel DAFormer as the backbone to separately compute features for dynamic and static objects during the training process. It also employs methods such as small category sampling and image blending to improve learning effectiveness. Additionally, we propose utilizing local image patches to enhance the results during the testing process. Experimental results demonstrate a significant improvement of 10.4% compared to the previous method, DANNet. Extensive experiments conducted on the Dark Zurich dataset and Nightdriving dataset validate the effectiveness of the proposed approach. Our method outperforms previous backbones and achieves top-ranking results.

Evaluating PVP coated iron oxide particles for localized magnetic hyperthermia and MRI imaging

Evaluating PVP coated iron oxide particles for localized magnetic hyperthermia and MRI imaging

A Metamaterial-like Structure Design Using Non-uniformly Distributed Dielectric and Conducting Strips to Boost the RF Field Distribution in 7 T MRI.

Metamaterial-based designs in ultra-high field (≥7 T) MRI have the promise of increasing the local magnetic resonance imaging (MRI) signal and potentially even the global efficiency of both the radiofrequency (RF) transmit and receive resonators. A recently proposed metamaterial-like structure-comprised of a high-permittivity dielectric material and a set of evenly distributed copper strips-indeed resulted in a local increase in RF transmission. Here, we demonstrate that non-uniform designs of this metamaterial-like structure can be used to boost the ultimate RF field distribution. A non-uniform dielectric distribution can yield longer electric dipoles, thus extending the RF transmit field coverage. A non-uniform distribution of conducting strips enables the tailoring of the local electric field hot spots, where a concave distribution resulted in lower power deposition. Simulations of the brain and calf regions using our new metamaterial-like design, which combines non-uniform distributions of both the dielectric and conducting strips, revealed a 1.4-fold increase in the RF field coverage compared to the uniform distribution, and a 1.5-2-fold increase in the transmit efficiency compared to the standard surface-coil.

국제 음악축제의 지역 애착도, 사전지식, 관여도가 지역 이미지, 축제에 대한 태도 및 지지도에 미치는 영향에 관한 연구

Purpose - This study examined the relationship between local attachment, prior knowledge and involvement in the festival, and local image, attitudes toward and support for the festival among people who have visited the Incheon Pentaport Music Festival, the largest music festival in Korea. Design/methodology/approach - A total of 250 samples were used for the final analysis, and SPSS 26.0 and AMOS 21.0 were used to analyze the data. The structural equation model analysis was conducted to test the hypotheses of this study. Findings - As a result of the analysis, first, only local attachment was found to have a positive and significant effect on local image. On the other hand, prior knowledge or involvement in the Incheon Pentaport Music Festival was found to be unrelated to local image. Second, local attachment and involvement were found to have a significant effect on attitude, but prior knowledge was not significant. Last, this paper examined the influence of festival area image and attitude toward the festival on festival support, and found that both have a positive and significant effect on support. Research implications or Originality - This study expands the scope of research on music festivals and is expected to contribute to tourist attraction and marketing strategies for the revitalization of music festivals in the future.

Empowering Local Image Generation: Harnessing Stable Diffusion for Machine Learning and AI

Empowering Local Image Generation: Harnessing Stable Diffusion for Machine Learning and AI

Environmental regulations and charitable donations made by polluting companies: Evidence from China

Using a quasi-natural experiment on the adjustment of sulfur dioxide emission levy standards by provinces in China from 2007 to 2014, we apply a difference-in-differences method to examine how stricter environmental regulations affect charitable donations made by polluting companies. Empirical results show that the increase in pollution levy standards increases corporate donations by 0.505%, implying that polluting companies use charitable donations as a strategic action to address the negative impact of environmental penalties. We also find that the political connection, local environmental governance preference, public image and corporate social responsibility strategies significantly affect the relationship between environmental regulations and donations of polluters.

TexFit: Text-Driven Fashion Image Editing with Diffusion Models

Fashion image editing aims to edit an input image to obtain richer or distinct visual clothing matching effects. Existing global fashion image editing methods are difficult to achieve rich outfit combination effects while local fashion image editing is more in line with the needs of diverse and personalized outfit matching. The local editing techniques typically depend on text and auxiliary modalities (e.g., human poses, human keypoints, garment sketches, etc.) for image manipulation, where the auxiliary modalities essentially assist in locating the editing region. Since these auxiliary modalities usually involve additional efforts in practical application scenarios, text-driven fashion image editing shows high flexibility. In this paper, we propose TexFit, a Text-driven Fashion image Editing method using diffusion models, which performs the local image editing only with the easily accessible text. Our approach employs a text-based editing region location module to predict precise editing region in the fashion image. Then, we take the predicted region as the generation condition of diffusion models together with the text prompt to achieve precise local editing of fashion images while keeping the rest part intact. In addition, previous fashion datasets usually focus on global description, lacking local descriptive information that can guide the precise local editing. Therefore, we develop a new DFMM-Spotlight dataset by using region extraction and attribute combination strategies. It focuses locally on clothes and accessories, enabling local editing with text input. Experimental results on the DFMM-Spotlight dataset demonstrate the effectiveness of our model. Code and Datasets are available at https://texfit.github.io/.

FeatWalk: Enhancing Few-Shot Classification through Local View Leveraging

Few-shot learning is a challenging task due to the limited availability of training samples. Recent few-shot learning studies with meta-learning and simple transfer learning methods have achieved promising performance. However, the feature extractor pre-trained with the upstream dataset may neglect the extraction of certain features which could be crucial for downstream tasks. In this study, inspired by the process of human learning in few-shot tasks, where humans not only observe the whole image (`global view') but also attend to various local image regions (`local view') for comprehensive understanding of detailed features, we propose a simple yet effective few-shot learning method called FeatWalk which can utilize the complementary nature of global and local views, therefore providing an intuitive and effective solution to the problem of insufficient local information extraction from the pre-trained feature extractor. Our method can be easily and flexibly combined with various existing methods, further enhancing few-shot learning performance. Extensive experiments on multiple benchmark datasets consistently demonstrate the effectiveness and versatility of our method.The source code is available at https://github.com/exceefind/FeatWalk.

CoPL: Contextual Prompt Learning for Vision-Language Understanding

Recent advances in multimodal learning has resulted in powerful vision-language models, whose representations are generalizable across a variety of downstream tasks. Recently, their generalization ability has been further extended by incorporating trainable prompts, borrowed from the natural language processing literature. While such prompt learning techniques have shown impressive results, we identify that these prompts are trained based on global image features which limits itself in two aspects: First, by using global features, these prompts could be focusing less on the discriminative foreground image, resulting in poor generalization to various out-of-distribution test cases. Second, existing work weights all prompts equally whereas intuitively, prompts should be reweighed according to the semantics of the image. We address these as part of our proposed Contextual Prompt Learning (CoPL) framework, capable of aligning the prompts to the localized features of the image. Our key innovations over earlier works include using local image features as part of the prompt learning process, and more crucially, learning to weight these prompts based on local features that are appropriate for the task at hand. This gives us dynamic prompts that are both aligned to local image features as well as aware of local contextual relationships. Our extensive set of experiments on a variety of standard and few-shot datasets show that our method produces substantially improved performance when compared to the current state of the art methods. We also demonstrate both few-shot and out-of-distribution performance to establish the utility of learning dynamic prompts that are aligned to local image features.