Visual Enhancement Research Articles

MVE-Net: A label-free microscopic image visual enhancement network via mRetinex and nonreference loss guidance

Label-free microscopic cell image analysis (segmentation, detection, counting, e.g.) is elementary for unravelling the biological functions of cells and their organelles. However, low contrast, darker brightness, background inhomogeneous, and weak edges of cells cause challenges in subsequent cell image analysis processes. To address these challenges, a Microscopic Visual Enhancement Network (MVE-Net) is proposed to improve microscopic visual effects through pre-enhancement and enhancement processes. In the pre-enhancement stage, to overcome the difficulty of acquiring paired or unpaired images for training, the mRetinex block is proposed to guide the pre-enhancement network to image contrast, cell details, and structural features. Furthermore, a multi-scale extraction module is employed to extract and fuse cell texture and structural features from the pre-enhanced images at various scales, guiding the generator training. In the enhancement stage, a nonreference loss block is designed, incorporating spatial consistency, uneven illumination smoothness, and exposure adjustment loss terms, to further enhance the contrast between cells and the background, smooth the inhomogeneous background, and adjust overall image brightness, thereby guiding the generator's enhancement process and improving the visual effect of microscopic images. Experiments on the LIVECell and PNT1A datasets demonstrate that MVE-Net outperforms state-of-the-art image enhancement methods, significantly improving image contrast, brightness, cell detail, and structural features without the need for paired or unpaired reference standard images for training.

Supporting Pharmacist-GP Collaboration in Medication Review Using Argumentation.

Medication review is a collaborative process between pharmacists and general practitioners (GPs) aimed at optimizing patient care by identifying and eliminating harmful medications. This paper proposes a collaborative platform to enhance pharmacist-GP interactions, assess drug-drug interactions, evaluate adverse effects, and manage dosages. The platform uses the issue mapping function of IBIS to structure dialogues and systematically evaluates proposed actions using the QuAD framework to support decision-making. An ontology based on medical knowledge ensures consistency, while visual enhancements such as varying edge width, color coding, and highlighting preferred actions enable swift, informed decisions. These tools improve collaboration and patient care outcomes.

Less gets more attention: A novel human-centered MR remote collaboration assembly method with information recommendation and visual enhancement

Mixed reality remote collaboration assembly is a type of computer-supported collaborative assembly work that uses mixed reality technology to enable spatial information and collaboration status sharing among geographically distributed collaborators, including remote experts and local users. However, due to the abundance of mixed virtual and real-world information in the MR space and the limitations imposed by narrow field-of-view augmented reality (AR) glasses, users face challenges in effectively focusing on relevant and valuable visual information. Our research aims to enhance users' visual attention to critical guidance information in MR collaborative assembly tasks, thereby improving the clear expression of instructions and facilitating the transmission of collaborative intention. We developed the Information Recommendation and Visual Enhancement System (IRVES) through an assembly process information hierarchy division mechanism, a content-based information recommendation system, and a gesture interaction-based information visual enhancement method. IRVES can leverage the guidance expertise and preferences of remote experts to recommend information to filter out irrelevant information and present the key information that the remote expert conveys to the local user in an intuitive way through visual enhancement. We conducted a user study experiment of a collaborative assembly task of a small engine in a laboratory environment. The experimental results indicate that IRVES outperforms traditional MR remote collaborative assembly methods (VG3DV) in terms of time performance, operational errors, cognitive performance and user experience. Our research contributes a human-centered information visualization approach for remote experts and local users, providing a novel method and idea for designing visual information interfaces in MR remote collaboration assembly tasks.

Conjugation of visual enhancers in lateral flow immunoassay for rapid forensic analysis: A critical review.

During crime scene investigations, numerous traces are secured and may be used as evidence for the evaluation of source and/or activity level propositions. The rapid chemical analysis of a biological trace enables the identification of body fluids and can provide significant donor profiling information, including age, sex, drug abuse, and lifestyle. Such information can be used to provide new leads, exclude from, or restrict the list of possible suspects during the investigative phase. This paper reviews the state-of-the-art labelling techniques to identify the most suitable visual enhancer to be implemented in a lateral flow immunoassay setup for the purpose of trace identification and/or donor profiling. Upon comparison, and with reference to the strengths and limitations of each label, the simplistic one-step analysis of noncompetitive lateral flow immunoassay (LFA) together with the implementation of carbon nanoparticles (CNPs) as visual enhancers is proposed for a sensitive, accurate, and reproducible in situ trace analysis. This approach is versatile and stable over different environmental conditions and external stimuli. The findings of the present comparative analysis may have important implications for future forensic practice. The selection of an appropriate enhancer is crucial for a well-designed LFA that can be implemented at the crime scene for a time- and cost-efficient investigation.

A Task-Guided, Implicitly-Searched and Meta-Initialized Deep Model for Image Fusion.

Image fusion plays a key role in a variety of multi-sensor-based vision systems, especially for enhancing visual quality and/or extracting aggregated features for perception. However, most existing methods just consider image fusion as an individual task, thus ignoring its underlying relationship with these downstream vision problems. Furthermore, designing proper fusion architectures often requires huge engineering labor. It also lacks mechanisms to improve the flexibility and generalization ability of current fusion approaches. To mitigate these issues, we establish a Task-guided, Implicit-searched and Meta-initialized (TIM) deep model to address the image fusion problem in a challenging real-world scenario. Specifically, we first propose a constrained strategy to incorporate information from downstream tasks to guide the unsupervised learning process of image fusion. Within this framework, we then design an implicit search scheme to automatically discover compact architectures for our fusion model with high efficiency. In addition, a pretext meta initialization technique is introduced to leverage divergence fusion data to support fast adaptation for different kinds of image fusion tasks. Qualitative and quantitative experimental results on different categories of image fusion problems and related downstream tasks (e.g., visual enhancement and semantic understanding) substantiate the flexibility and effectiveness of our TIM.

Birdlike broadband neuromorphic visual sensor arrays for fusion imaging.

Wearable visual bionic devices, fueled by advancements in artificial intelligence, are making remarkable progress. However, traditional silicon vision chips often grapple with high energy losses and challenges in emulating complex biological behaviors. In this study, we constructed a van der Waals P3HT/GaAs nanowires P-N junction by carefully directing the arrangement of organic molecules. Combined with a Schottky junction, this facilitated multi-faceted birdlike visual enhancement, including broadband non-volatile storage, low-light perception, and a near-zero power consumption operating mode in both individual devices and 5 × 5 arrays on arbitrary substrates. Specifically, we realized over 5 bits of in-memory sensing and computing with both negative and positive photoconductivity. When paired with two imaging modes (visible and UV), our reservoir computing system demonstrated up to 94% accuracy for color recognition. It achieved motion and UV grayscale information extraction (displayed with sunscreen), leading to fusion visual imaging. This work provides a promising co-design of material and device for a broadband and highly biomimetic optoelectronic neuromorphic system.

Spectrum aided vision enhancer enhances mucosal visualization by hyperspectral imaging in capsule endoscopy.

Narrow-band imaging (NBI) is more efficient in detecting early gastrointestinal cancer than white light imaging (WLI). NBI technology is available only in conventional endoscopy, but unavailable in magnetic-assisted capsule endoscopy (MACE) systems due to MACE's small size and obstacles in image processing issues. MACE is an easy, safe, and convenient tool for both patients and physicians to avoid the disadvantages of conventional endoscopy. Enabling NBI technology in MACE is mandatory. We developed a novel method to improve mucosal visualization using hyperspectral imaging (HSI) known as Spectrum Aided Visual Enhancer (SAVE, Transfer N,Hitspectra Intelligent Technology Co., Kaohsiung, Taiwan). The technique was developed by converting the WLI image captured by MACE to enhance SAVE images. The structural similarity index metric (SSIM) between the WLI MACE images and the enhanced SAVE images was 91%, while the entropy difference between the WLI MACE images and the enhanced SAVE images was only 0.47%. SAVE algorithm can identify the mucosal break on the esophagogastric junction in patients with gastroesophageal reflux disorder. We successfully developed a novel image-enhancing technique, SAVE, in the MACE system, showing close similarity to the NBI from the conventional endoscopy system. The future application of this novel technology in the MACE system can be promising.

Modernizing Public Health Data Systems and Workforce Capacity: The Centers for Disease Control and Prevention's Public Health Informatics Fellowship Program.

The COVID-19 pandemic exposed governmental public health's outdated information technology and insufficient data science and informatics workforce capacity. The Centers for Disease Control and Prevention's Public Health Informatics Fellowship Program (PHIFP) is well positioned to strengthen public health data science and informatics workforce capacity. Established in 1996, PHIFP is a 2-year, full-time, on-the-job training program. PHIFP includes a didactic curriculum, applied learning through informatics projects completed at the assigned host site, short-term technical assistance projects, and a final capstone project. Fellows have learned from and bolstered host site informatics capacity through the development or enhancement of information systems, evaluations, data integration, data visualization, and analysis. Among recent graduates, 54% are employed at Centers for Disease Control and Prevention and 16% are employed at other public health organizations, including local health departments. Fellowships such as PHIFP, which recruit and train promising scientists in public health informatics, are important components of efforts to strengthen public health workforce capacity.

MA-VoxelMorph: Multi-scale attention-based VoxelMorph for nonrigid registration of thoracoabdominal CT images

This paper aims to develop a nonrigid registration method of preoperative and intraoperative thoracoabdominal CT images in computer-assisted interventional surgeries for accurate tumor localization and tissue visualization enhancement. However, fine structure registration of complex thoracoabdominal organs and large deformation registration caused by respiratory motion is challenging. To deal with this problem, we propose a 3D multi-scale attention VoxelMorph (MA-VoxelMorph) registration network. To alleviate the large deformation problem, a multi-scale axial attention mechanism is utilized by using a residual dilated pyramid pooling for multi-scale feature extraction, and position-aware axial attention for long-distance dependencies between pixels capture. To further improve the large deformation and fine structure registration results, a multi-scale context channel attention mechanism is employed utilizing content information via adjacent encoding layers. Our method was evaluated on four public lung datasets (DIR-Lab dataset, Creatis dataset, Learn2Reg dataset, OASIS dataset) and a local dataset. Results proved that the proposed method achieved better registration performance than current state-of-the-art methods, especially in handling the registration of large deformations and fine structures. It also proved to be fast in 3D image registration, using about 1.5 s, and faster than most methods. Qualitative and quantitative assessments proved that the proposed MA-VoxelMorph has the potential to realize precise and fast tumor localization in clinical interventional surgeries.

Comparing macular thickness and visual acuity responses retinal vein occlusion versus diabetic macular edema following anti vascular endothelial growth factor therapy: A retrospective analysis

: Vascular growth factor associates Retinal vein occlusion (RVO) and Diabetic Macular Edema (DME). A retrospective analysis using database from a tertiary hospital.: The study aims to evaluate effects of anti-vascular growth factor therapy on patients Diabetic Macular Edema with and retinal vein occlusion,: This study conducted between December 2022 and June 2023 where 30 patients diagnosed with Diabetic macular edema (DME, n=15) or Retinal vein occlusion (RVO, n=15) receiving anti-VEGF were examined. This focused on changes in best corrected visual acuity, central macular thickness and predictors over a period of three months.: Retinal Vein Occlusion (RVO) patients exhibited a more significant improvement in Best Corrected Visual Acuity (BCVA) compared to Diabetic Macular Edema (DME) (25 vs. 10 letters, p = 0.006) after three months. The change in Central Macular Thickness (CMT) also favored RVO (100µm) over DME (40µm, p = 0.012). Final BCVA median was 70 for DME and 65 for RVO (p = 0.461). Considering initial BCVA and CMT, RVO predicted better visual enhancement than DME.: RVO exhibited major BCVA improvement at three months with Anti-Vascular Endothelial Growth Factor. Final vision was indistinguishable for DME and RVO. Anti VEGF stabilizes DME vision. Despite the improvement, early RVO diagnosis and treatment may enhance final vision.

Visual Enhancement of Aerial Flight in High Plateau Low Visibility Conditions

A method for automatic detection of airport runways in low visibility, forward and downward looking aerial airport images is proposed. Firstly, the recursive Otsu segmentation method is used to extract the main contour of the airport from the complex background; then the morphological filtering is performed by using the morphology-based open and swell operations, the open operation is used to filter out the isolated noise points and the swell operation is used to connect the segmented airport components into a complete whole; finally, the morphological filtered image is subjected to the region growth to localize the main runway of the airport. Finally, the morphology filtered image is processed to localize the main airport runway by its region growth. It is proved that this method can quickly and accurately detect and localize the main and auxiliary runways of airports in low visibility and forward and downward looking aerial images of airports.

Interactive dual-stream contrastive learning for radiology report generation

Radiology report generation automates diagnostic narrative synthesis from medical imaging data. Current report generation methods primarily employ knowledge graphs for image enhancement, neglecting the interpretability and guiding function of the knowledge graphs themselves. Additionally, few approaches leverage the stable modal alignment information from multimodal pre-trained models to facilitate the generation of radiology reports. We propose the Terms-Guided Radiology Report Generation (TGR), a simple and practical model for generating reports guided primarily by anatomical terms. Specifically, we utilize a dual-stream visual feature extraction module comprised of detail extraction module and a frozen multimodal pre-trained model to separately extract visual detail features and semantic features. Furthermore, a Visual Enhancement Module (VEM) is proposed to further enrich the visual features, thereby facilitating the generation of a list of anatomical terms. We integrate anatomical terms with image features and proceed to engage contrastive learning with frozen text embeddings, utilizing the stable feature space from these embeddings to boost modal alignment capabilities further. Our model incorporates the capability for manual input, enabling it to generate a list of organs for specifically focused abnormal areas or to produce more accurate single-sentence descriptions based on selected anatomical terms. Comprehensive experiments demonstrate the effectiveness of our method in report generation tasks, our TGR-S model reduces training parameters by 38.9% while performing comparably to current state-of-the-art models, and our TGR-B model exceeds the best baseline models across multiple metrics.

Visual object detection system based on augmented reality and steady-state visual evoked potential

This study investigates a brain-computer interface (BCI) system based on an augmented reality (AR) environment and steady-state visual evoked potentials (SSVEP). The system is designed to facilitate the selection of real-world objects through visual gaze in real-life scenarios. By integrating object detection technology and AR technology, the system augmented real objects with visual enhancements, providing users with visual stimuli that induced corresponding brain signals. SSVEP technology was then utilized to interpret these brain signals and identify the objects that users focused on. Additionally, an adaptive dynamic time-window-based filter bank canonical correlation analysis was employed to rapidly parse the subjects' brain signals. Experimental results indicated that the system could effectively recognize SSVEP signals, achieving an average accuracy rate of 90.6% in visual target identification. This system extends the application of SSVEP signals to real-life scenarios, demonstrating feasibility and efficacy in assisting individuals with mobility impairments and physical disabilities in object selection tasks.

FastUNet: Fast hierarchical multi-patch underwater enhancement network for industrial recirculating aquaculture

The industrialized aquaculture with recirculating water systems has gained increasing attention and rapid development in recent years. However, the existing underwater enhancement methods are time-consuming and severely hinder their online application in industrialized aquaculture due to the requirement for real-time processing. We found that their main limitations are the insufficient utilization of the hierarchical network approach and valuable prior information in underwater scenarios. To address this problem, we propose a novel architecture called FastUNet and introduce a visual enhancement benchmark dataset specifically designed for industrialized aquaculture with recirculating water systems. Surprisingly, by learning the fast multi-patch hierarchical module and employing prior constraints including color loss and homology loss, FastUNet achieves comparable or even better results than many deep learning methods. In terms of runtime, it is up to 30 times faster than current underwater enhancement methods. The processing time for images with a resolution of 5474 × 3653 is only 0.137s. The real application scenario of factory recirculating aquaculture is considered by FastUNet, and the embedding of prior information under the fast network framework is explored. In this paper, 14 state-of-the-art underwater enhancement methods are compared, and excellent results are obtained on 2 Full-reference datasets and 2 No-referenced datasets. The application test results of SIFT significant point detection, geometric rotation, edge detection, target segmentation and target detection were also presented, which strongly verified the feasibility of FastUNet in the field of factory recirculating aquaculture.

Decoupled Cross-Modal Transformer for Referring Video Object Segmentation.

Referring video object segmentation (R-VOS) is a fundamental vision-language task which aims to segment the target referred by language expression in all video frames. Existing query-based R-VOS methods have conducted in-depth exploration of the interaction and alignment between visual and linguistic features but fail to transfer the information of the two modalities to the query vector with balanced intensities. Furthermore, most of the traditional approaches suffer from severe information loss in the process of multi-scale feature fusion, resulting in inaccurate segmentation. In this paper, we propose DCT, an end-to-end decoupled cross-modal transformer for referring video object segmentation, to better utilize multi-modal and multi-scale information. Specifically, we first design a Language-Guided Visual Enhancement Module (LGVE) to transmit discriminative linguistic information to visual features of all levels, performing an initial filtering of irrelevant background regions. Then, we propose a decoupled transformer decoder, using a set of object queries to gather entity-related information from both visual and linguistic features independently, mitigating the attention bias caused by feature size differences. Finally, the Cross-layer Feature Pyramid Network (CFPN) is introduced to preserve more visual details by establishing direct cross-layer communication. Extensive experiments have been carried out on A2D-Sentences, JHMDB-Sentences and Ref-Youtube-VOS. The results show that DCT achieves competitive segmentation accuracy compared with the state-of-the-art methods.

Assessing the Efficacy of the Spectrum-Aided Vision Enhancer (SAVE) to Detect Acral Lentiginous Melanoma, Melanoma In Situ, Nodular Melanoma, and Superficial Spreading Melanoma.

Skin cancer is the predominant form of cancer worldwide, including 75% of all cancer cases. This study aims to evaluate the effectiveness of the spectrum-aided visual enhancer (SAVE) in detecting skin cancer. This paper presents the development of a novel algorithm for snapshot hyperspectral conversion, capable of converting RGB images into hyperspectral images (HSI). The integration of band selection with HSI has facilitated the identification of a set of narrow band images (NBI) from the RGB images. This study utilizes various iterations of the You Only Look Once (YOLO) machine learning (ML) framework to assess the precision, recall, and mean average precision in the detection of skin cancer. YOLO is commonly preferred in medical diagnostics due to its real-time processing speed and accuracy, which are essential for delivering effective and efficient patient care. The precision, recall, and mean average precision (mAP) of the SAVE images show a notable enhancement in comparison to the RGB images. This work has the potential to greatly enhance the efficiency of skin cancer detection, as well as improve early detection rates and diagnostic accuracy. Consequently, it may lead to a reduction in both morbidity and mortality rates.

Advancements in needle visualization enhancement and localization methods in ultrasound: a literature review

Ultrasound guidance plays a central role in numerous minimally invasive procedures involving percutaneous needle insertion, ensuring safe and accurate needle placement. However, it encounters two primary challenges: (1) aligning the needle with the ultrasound beam and (2) visualizing the needle even when correctly aligned. In this review, we offer a concise overview of the physics foundation underlying these challenges and explore various approaches addressing specific challenges, with a focus on software-based solutions. We further distinguish between hardware-based and software-based solutions, placing a stronger emphasis on the latter. The incorporation of artificial intelligence into these methods to enhance needle visualization and localization is briefly discussed. We identify state-of-the-art needle detection methods, showcasing submillimeter precision in tip localization and orientation. Additionally, we provide insights into potential future directions, aiming to facilitate the translation of these advanced methods into the clinic. This article serves as a comprehensive guide, offering insights into challenges, evolving solutions, and prospective research directions to effectively address these issues.

Integration of two different road network models for emergency rescue pathfinding in indoor and outdoor environments

Abstract The environment of buildings and outdoor transportation in urban areas is becoming increasingly complex. This poses a challenge for rescue pathfinding once an emergency occurs, which relies on the generation of road network models. To better balance efficiency and accuracy in pathfinding, a three-dimensional integrated road network model (3D-IRNM) derived from grid and topological road network models was developed in this study. Firstly, a cross-platform data loading method from building information model to geographic information systems (GIS) is proposed, to serve as a data source for road network model generation, visual enhancement, and breaking the barriers between different spatial data format models. Then, the topological and grid road network models are generated in indoor spaces of different functions, with IRNM obtained by the integration strategy. The 3D-IRNM can be formed by extracting indoor vertical paths from the stairs and connecting them with the IRNM of each floor. To further combine the 3D-IRNM with outdoor environments, a CrossNode model is then proposed. Besides, an adaptive pathfinding algorithm is also proposed. Finally, the construction of an emergency pathfinding system based on GIS technology is achieved. The pathfinding algorithm and the structure of 3D-IRNM help the pathfinding both efficiency and accuracy. The practicability of the designed 3D-IRNM together with its pathfinding algorithm is well verified.

Improving GPR interpretation by applying image-enhancing attributes and machine learning techniques: A case study over Green Hill Cemetery in Frankfort, Kentucky

Ground-penetrating radar (GPR) enables noninvasive imaging in structural mapping applications such as unmarked grave detection. However, burial signatures can be cryptic and require expert analysis. This study investigates attribute enhancement and machine learning to automate identification of variable-signature graves at Green Hill Cemetery in Frankfort, Kentucky. Nine complementary seismic attributes were computed from the GPR envelope reflectivity volume to boost discontinuities and patterns indicative of shafts. Coherent energy, pseudofrequency, and similarity transforms showed optimal visualization enhancement. These volumes were input into unsupervised k-means and self-organizing map (SOM) machine learning models to cluster potential burial sites. Both methods accurately characterized strong anomaly reflections associated with apparent grave boundaries. However, limitations emerged in classifying subtler signatures that are likely linked to deteriorated or deeper interments. SOM clustering provided finer segmentation between noise and targets. Collectively, attributes amplified burial edges for easier recognition, while machine learning clustering expedited identification of most vault structures and unmarked sites. However, edge case discrimination remains a challenge. Results suggest that hybrid human-machine learning analysis can enhance efficiency over purely manual interpretation. With further method refinement, automated attribute and machine learning workflows show strong potential for accelerating GPR-based cemetery and archaeological mapping.

Autonomous underwater vehicle visual enhancement using area-to-point kriging and multi-color spaces embedding

Autonomous underwater vehicle visual enhancement using area-to-point kriging and multi-color spaces embedding