High-quality Images Research Articles

Face Image Generation for Anime Characters based on Generative Adversarial Network

With the increasing demand for digital art, animation, and games, facial generation for anime characters has attracted growing research interest in recent years, which aims to build models to automatically generate unique and high-quality character images. Thanks to the rapid advancement of deep learning techniques, particularly generative adversarial networks, GAN-based image generation methods have continuously achieved breakthroughs in generation effectiveness and speed. Focusing on generating realistic anime face images, this paper proposes an anime character face image generation model based on GANs, which integrates Bath Normalization and Dropout to maintain strong stability and avoid overfitting. Comprehensive experiments show the efficacy of the proposed method, which can achieve high diversity in facial features and styles while maintaining visual coherence

Artificial Intelligence-Guided Lung Ultrasound by Nonexperts.

Lung ultrasound (LUS) aids in the diagnosis of patients with dyspnea, including those with cardiogenic pulmonary edema, but requires technical proficiency for image acquisition. Previous research has demonstrated the effectiveness of artificial intelligence (AI) in guiding novice users to acquire high-quality cardiac ultrasound images, suggesting its potential for broader use in LUS. To evaluate the ability of AI to guide acquisition of diagnostic-quality LUS images by trained health care professionals (THCPs). In this multicenter diagnostic validation study conducted between July 2023 and December 2023, participants aged 21 years or older with shortness of breath recruited from 4 clinical sites underwent 2 ultrasound examinations: 1 examination by a THCP operator using Lung Guidance AI and the other by a trained LUS expert without AI. The THCPs (including medical assistants, respiratory therapists, and nurses) underwent standardized AI training for LUS acquisition before participation. Lung Guidance AI software uses deep learning algorithms guiding LUS image acquisition and B-line annotation. Using an 8-zone LUS protocol, the AI software automatically captures images of diagnostic quality. The primary end point was the proportion of THCP-acquired examinations of diagnostic quality according to a panel of 5 masked expert LUS readers, who provided remote review and ground truth validation. The intention-to-treat analysis included 176 participants (81 female participants [46.0%]; mean [SD] age, 63 [14] years; mean [SD] body mass index, 31 [8]). Overall, 98.3% (95% CI, 95.1%-99.4%) of THCP-acquired studies were of diagnostic quality, with no statistically significant difference in quality compared to LUS expert-acquired studies (difference, 1.7%; 95% CI, -1.6% to 5.0%). In this multicenter validation study, THCPs with AI assistance achieved LUS images meeting diagnostic standards compared with LUS experts without AI. This technology could extend access to LUS to underserved areas lacking expert personnel. ClinicalTrials.gov Identifier: NCT05992324.

An Investigation of Four-Point Bending Behavior for Personalized Humerus Bone Plate

The reverse engineering method has been beneficial for personalized implant designs in the medical field due to its flexible use feature that allows it to be applied in many areas. The method of anatomical features (MAF) is commonly used for fractures of the bone for the development of plate fixation. To improve the geometrical fitting between plate and bone, high-quality image data is the most important factor for computer aided design (CAD ) modelling. This study aims to use MAF for humerus bone plate design that would alternate the fracture region and decrease the stress-shielding effect depending on bending strength. For this purpose, the personalized plate implant (PPI) was designed according to MAF and fabricated using the multi jet fusion (MJF) technique, finally the four-point bending finite element analysis was applied and tested. The results indicated that the PPI structure has higher bending strength than the flat plate design and a much greater surface area.

PREmbed: Balancing Conditional Generative Models with Embedding Pretraining and Regularization

Recent advancements in conditional generative models, e.g., Conditional Variational AutoEncoder (CVAE) and Conditional Denoising Diffusion Probabilistic Model (CDDPM), which utilize class-specific embeddings for generating images in specific classes, have demonstrated exceptional capabilities in producing high-quality images on balanced datasets. However, their performance decreases with imbalanced real-world data, affecting the fidelity and diversity of the generated images. This paper discusses the reasons and solutions for the imbalance issue in CVAE and CDDPM. By selectively reweighting the gradients of the embedding layer and main network, we identify the embedding layer’s bias as a critical bottleneck in these models. Motivated by this, we propose Embedding Pretraining and Regularization (PREmbed), a training methodology for addressing biased learning in the embedding layer of these models. PREmbed consists of two key components: (i) Masked Autoencoder pretraining on the original imbalanced dataset to obtain a robust initial embedding, and (ii) an embedding regularization loss to preserve class-level distances during the training phase to improve imbalanced embedding learning. Implementing PREmbed leads to the enhanced convergence of a robust embedding layer. Our experiments on imbalanced CIFAR-10 and CUB-200 datasets demonstrate that PREmbed enhances the performance of CVAE and CDDPM; it outperforms the baseline CVAE by 22.3% (FID 18.63) and the CDDPM by 30.3% (FID 12.70) in imbalanced CIFAR-10; 33.7% (FID 24.08) and 26.5% (FID 13.90) in CUB-200; and 16.4% (FID 49.41) and 15.2% (FID 33.05) in ImageNet-LT (which is where it achieves the state-of-the-art performance in imbalanced conditional image generation task).

Human Identification System using Images

The present study focuses on the challenges of human long-range recognition with deep learning, and AlexNet performance is considered, being one of the most recognized convolutional neural networks. A specially curated dataset was used to mimic the real world in a more practical and realistic scenario, which involves subjects varying in distances and angles while creating low-resolution issues with variability in lighting and distractions from the environment. The system starts by passing the input dataset through a local Laplacian filter that is utilized to enhance the quality of images before feeding it into a pre-trained AlexNet model. The results reveal that AlexNet reaches a recognition accuracy of 65% and shows that image properties such as sharpness and contrast significantly affect performance, as measured by the mean and standard deviation. The study underlines requirements of high-quality images for applications in long-range human recognition and sheds insight into the ways enhancements in images could improve performance in deep learning models. Results support further evolution of more potent human recognition systems, particularly in application domains such as surveillance and security in which the accurate identification of a target from a long distance is highly critical. Key Words: Long-range recognition, Human recognition, Deep learning, Image processing

Deep learning multi-classification of middle ear diseases using synthetic tympanic images

Background Recent advances in artificial intelligence have facilitated the automatic diagnosis of middle ear diseases using endoscopic tympanic membrane imaging. Aim We aimed to develop an automated diagnostic system for middle ear diseases by applying deep learning techniques to tympanic membrane images obtained during routine clinical practice. Material and methods To augment the training dataset, we explored the use of generative adversarial networks (GANs) to produce high-quality synthetic tympanic images that were subsequently added to the training data. Between 2016 and 2021, we collected 472 endoscopic images representing four tympanic membrane conditions: normal, acute otitis media, otitis media with effusion, and chronic suppurative otitis media. These images were utilized for machine learning based on the InceptionV3 model, which was pretrained on ImageNet. Additionally, 200 synthetic images generated using StyleGAN3 and considered appropriate for each disease category were incorporated for retraining. Results The inclusion of synthetic images alongside real endoscopic images did not significantly improve the diagnostic accuracy compared to training solely with real images. However, when trained solely on synthetic images, the model achieved a diagnostic accuracy of approximately 70%. Conclusions and significance Synthetic images generated by GANs have potential utility in the development of machine-learning models for medical diagnosis.

The molar dose of FAPI administered impacts on the FAP-targeted PET imaging and therapy in mouse syngeneic tumor models.

Since fibroblast activation protein (FAP), one predominant biomarker of cancer associated fibroblasts (CAFs), is highly expressed in the tumor stroma of various epidermal-derived cancers, targeting FAP for tumor diagnosis and treatment has shown substantial potentials in both preclinical and clinical studies. However, in preclinical settings, tumor-bearing mice exhibit relatively low absolute FAP expression levels, leading to challenges in acquiring high-quality PET images using radiolabeled FAP ligands (FAPIs) with low molar activity, because of which a saturation effect in imaging is prone to happen. Moreover, how exactly the molar dose of FAPI administered to a mouse influences the targeted PET imaging and radiotherapy remains unclear now. Therefore, this study aims to investigate the impacts of the molar dose of the administered FAPI on FAP-targeted PET imaging and radiotherapy in mouse syngeneic tumor models. [68Ga]Ga-FAPI-04 with various molar doses of FAPI-04 was administered to wild-type 4T1 tumor-bearing mice, followed by static PET imaging. Sigmoidal curves were generated to analyze the correlation between the standard uptake value (SUV) and the administered molar doses of FAPI-04. Similarly, [177Lu]Lu-DOTAGA.(SA.FAPi)2 with a consistent dose of radioactivity but containing different moles of DOTAGA.(SA.FAPi)2 were injected into 4T1 tumor-bearing mice to assess the therapeutic effect. [68Ga]Ga-FAPI-04 was also applied to different tumor models for PET/CT imaging. A gradient blocking effect was observed with increasing FAPI molar dose in [68Ga]Ga-FAPI-04 PET imaging and [177Lu]Lu-DOTAGA.(SA.FAPi)2 treatment, with various imaging and therapeutic outcomes. [68Ga]Ga-FAPI-04 PET exhibit potentials to characterize murine derived FAP expression with low molar dose of administered FAPI-04 using various tumor models. The molar dose of FAPI in [68Ga]Ga/[177Lu]Lu-FAPI had a substantial impact on FAP-targeted imaging and therapy in mouse syngeneic tumor models. To acquire enhanced reliability and reproducibility in preclinical situation, it is critical to carefully consider the molar dose of the radiotracer when applying radiolabeled FAP ligands to FAP-targeted imaging and radiotherapy.

High-performance solar-blind ultraviolet photodetector arrays based on two-inch ϵ-Ga2O3 films for imaging applications

Abstract To achieve high-quality solar-blind imaging applications based on ultrawide bandgap semiconductor photodetectors, it is crucial to fabricate highly uniform wafer-scale films. In this work, we demonstrate the fabrication of exceptionally uniform two-inch ε-Ga2O3 thin films on sapphire substrates using an off-axis pulsed laser deposition method. The two-inch ε-Ga2O3 films exhibit remarkable uniformity across key parameters, including thickness, crystalline quality, bandgap, and surface roughness, with an inhomogeneity ratio less than 5%. Additionally, these films are preferentially oriented along the (001) crystal plane. At 20 V bias, the individual ε-Ga2O3 photodetector demonstrates outstanding solar-blind ultraviolet (UV) photodetection performance, with a responsivity of 52.77 A/W at 240 nm, an external quantum efficiency of 2.7×104%, a dark current of 5.5×10-11 A and a UV–visible rejection ratio of 1.2×104. Furthermore, the 10×10 photodetector arrays fabricated on two-inch ε-Ga2O3 films exhibit highly uniform photodetection performance, with photocurrent deviations remaining within one order of magnitude and a maximum standard deviation of ~8%. High-contrast optical imaging of the letters of “NIMTE” is successfully achieved using the 10×10 photodetector array detector. This work provides valuable insights for fabricating wafer-scale uniform ε-Ga2O3 films and achieving high-quality solar-blind UV imaging applications.

Multi-wavelength global maps of Jupiter and Saturn using Cassini imaging system data

The Imaging Science Subsystem onboard the Cassini spacecraft recorded numerous high-quality images of Jupiter and Saturn at various wavelengths, from ultraviolet to near-infrared, during its 20-year mission from 1997 to 2017. Using these images, we have developed global maps of Jupiter and Saturn across multiple wavelengths. These maps reveal the global atmospheric structures of Jupiter and Saturn, offering a comprehensive tool to study the physical and dynamic processes of these atmospheric systems on a global scale. Additionally, these multi-wavelength maps, which probe different pressure levels within the atmospheres, help to explore the vertical structure of these processes. Moreover, global maps at different times enable tracking the movement of dynamic phenomena (e.g., clouds, storms, vortices, eddies, waves, and turbulence), thereby enhancing our understanding of atmospheric dynamics on the giant planets.

Multiparametric MRI for Assessment of the Biological Invasiveness and Prognosis of Pancreatic Ductal Adenocarcinoma in the Era of Artificial Intelligence.

Pancreatic ductal adenocarcinoma (PDAC) is the deadliest malignant tumor, with a grim 5-year overall survival rate of about 12%. As its incidence and mortality rates rise, it is likely to become the second-leading cause of cancer-related death. The radiological assessment determined the stage and management of PDAC. However, it is a highly heterogeneous disease with the complexity of the tumor microenvironment, and it is challenging to adequately reflect the biological aggressiveness and prognosis accurately through morphological evaluation alone. With the dramatic development of artificial intelligence (AI), multiparametric magnetic resonance imaging (mpMRI) using specific contrast media and special techniques can provide morphological and functional information with high image quality and become a powerful tool in quantifying intratumor characteristics. Besides, AI has been widespread in the field of medical imaging analysis. Radiomics is the high-throughput mining of quantitative image features from medical imaging that enables data to be extracted and applied for better decision support. Deep learning is a subset of artificial neural network algorithms that can automatically learn feature representations from data. AI-enabled imaging biomarkers of mpMRI have enormous promise to bridge the gap between medical imaging and personalized medicine and demonstrate huge advantages in predicting biological characteristics and the prognosis of PDAC. However, current AI-based models of PDAC operate mainly in the realm of a single modality with a relatively small sample size, and the technical reproducibility and biological interpretation present a barrage of new potential challenges. In the future, the integration of multi-omics data, such as radiomics and genomics, alongside the establishment of standardized analytical frameworks will provide opportunities to increase the robustness and interpretability of AI-enabled image biomarkers and bring these biomarkers closer to clinical practice. EVIDENCE LEVEL: 3 TECHNICAL EFFICACY: Stage 4.

Simultaneous Luminal and Hemodynamic Evaluation of the Cervical Arteries Using Nonenhanced 3D Quantitative Quiescent-Interval Slice-Selective Magnetic Resonance Angiography.

Luminal and hemodynamic evaluations of the cervical arteries inform the diagnosis and management of patients with cervical arterial disease. To demonstrate a 3D nonenhanced quantitative quiescent interval slice-selective (qQISS) magnetic resonance angiographic (MRA) strategy that provides simultaneous hemodynamic and luminal evaluation of the cervical arteries. Prospective. Six healthy volunteers (3 female, 3 male, age = 35.7 ± 10.3 years) and 14 patients with cerebrovascular disease (12 female, 2 male, age = 56.6 ± 14.0 years). 3 T, ungated 3D tilted-slab qQISS, pulse-gated 2D phase contrast (PC), ungated 3D PC, and 3D time-of-flight (TOF)gradient-echo protocols. Four readers scored 29 arterial segments on 3D qQISS volumes for image quality using a 4-point scale (1: non-diagnostic, 2: fair, 3: good, 4: excellent). Time-averaged arterial flow velocities and volume flow rates obtained with qQISS and PC protocols were compared. Arterial lumen area and radius measures obtained with 3D protocols were compared in a subgroup. Gwet's AC2; intraclass correlation coefficient (ICC); Pearson's correlation; Bland-Altman. P values <0.05 were considered statistically significant. 3D qQISS provided good-to-excellent image quality for depicting the cervical arteries (mean scores of 3.72 ± 0.55, 3.55 ± 0.66, 3.42 ± 0.72, and 3.66 ± 0.73 for readers 1, 2, 3, and 4) with significant inter-reader agreement (AC2 = 0.91, ICC = 0.53) in image scoring, significantly agreed with pulse-gated 2D PC for time-averaged total flow velocity (ICC = 0.83) and volume flow rate (ICC = 0.92), and significantly agreed with 3D PC for total flow velocity (ICC = 0.70), volume flow rate (ICC = 0.91), and component flow velocity (ICC = 0.89). Compared with 3D PC, 3D qQISS better agreed with 3D TOF for arterial lumen area (ICC = 0.97 vs. 0.72) and radius (ICC = 0.94 vs. 0.74). Nonenhanced 3D qQISS provides high-quality sub-1 mm3 spatial resolution imaging of the cervical arteries, excellent agreement of arterial structural measures with respect to 3D TOF, and time-averaged hemodynamic data without the need for additional PC imaging. Magnetic resonance angiography (MRA), a method for depicting blood vessels within the body, can be used to evaluate arterial diseases and disorders of the neck. MRA methods routinely used to evaluate the neck arteries do not measure blood flow speed and volume, while other methods for obtaining this information provide less accurate pictures of arterial structure and are not routinely collected. This article reports a new method for MRA that clearly and efficiently portrays the neck arteries without using injected dyes, and provides measurements of arterial blood flow speed and volume. 2 TECHNICAL EFFICACY: Stage 1.

Robust broadband ptychography algorithms for high-harmonic soft X-ray supercontinua

We demonstrate PaCMAN, a ptychography algorithm that can reconstruct high quality images with broadband illumination sources while being robust to shot, detector, and parasitic noise. We extend prior monochromatization work to improve accuracy, especially for discrete spectra, and also demonstrate how PaCMAN can be converted into Ms. PaCMAN, a multi-spectral variant that outperforms multi-spectral ePIE. When comparing speed-optimized ePIE and PaCMAN, we achieve high reconstruction quality in 4x less time, and when they are dose-optimized, we achieve high reconstruction quality in 1.5-2x less dose. These algorithms will enable rapid sub-10 nm imaging with table-top soft X-ray high harmonic supercontinua, advance attosecond coherent diffractive imaging, and reduce the dose needed to image radiation sensitive materials.

Automatic image generation and stage prediction of breast cancer immunobiological through a proposed IHC-GAN model.

Invasive breast cancer diagnosis and treatment planning require an accurate assessment of human epidermal growth factor receptor 2 (HER2) expression levels. While immunohistochemical techniques (IHC) are the gold standard for HER2 evaluation, their implementation can be resource-intensive and costly. To reduce these obstacles and expedite the procedure, we present an efficient deep-learning model that generates high-quality IHC-stained images directly from Hematoxylin and Eosin (H&E) stained images. We propose a new IHC-GAN that enhances the Pix2PixHD model into a dual generator module, improving its performance and simplifying its structure. Furthermore, to strengthen feature extraction for HE-stained image classification, we integrate MobileNetV3 as the backbone network. The extracted features are then merged with those generated by the generator to improve overall performance. Moreover, the decoder's performance is enhanced by providing the related features from the classified labels by incorporating the adaptive instance normalization technique. The proposed IHC-GAN was trained and validated on a comprehensive dataset comprising 4,870 registered image pairs, encompassing a spectrum of HER2 expression levels. Our findings demonstrate promising results in translating H&E images to IHC-equivalent representations, offering a potential solution to reduce the costs associated with traditional HER2 assessment methods. We extensively validate our model and the current dataset. We compare it with state-of-the-art techniques, achieving high performance using different evaluation metrics, showing 0.0927 FID, 22.87 PSNR, and 0.3735 SSIM. The proposed approach exhibits significant enhancements over current GAN models, including an 88% reduction in Frechet Inception Distance (FID), a 4% enhancement in Learned Perceptual Image Patch Similarity (LPIPS), a 10% increase in Peak Signal-to-Noise Ratio (PSNR), and a 45% reduction in Mean Squared Error (MSE). This advancement holds significant potential for enhancing efficiency, reducing manpower requirements, and facilitating timely treatment decisions in breast cancer care.

A phase transition in diffusion models reveals the hierarchical nature of data

Understanding the structure of real data is paramount in advancing modern deep-learning methodologies. Natural data such as images are believed to be composed of features organized in a hierarchical and combinatorial manner, which neural networks capture during learning. Recent advancements show that diffusion models can generate high-quality images, hinting at their ability to capture this underlying compositional structure. We study this phenomenon in a hierarchical generative model of data. We find that the backward diffusion process acting after a time t is governed by a phase transition at some threshold time, where the probability of reconstructing high-level features, like the class of an image, suddenly drops. Instead, the reconstruction of low-level features, such as specific details of an image, evolves smoothly across the whole diffusion process. This result implies that at times beyond the transition, the class has changed, but the generated sample may still be composed of low-level elements of the initial image. We validate these theoretical insights through numerical experiments on class-unconditional ImageNet diffusion models. Our analysis characterizes the relationship between time and scale in diffusion models and puts forward generative models as powerful tools to model combinatorial data properties.

Diffusion data augmentation for enhancing Norberg hip angle estimation.

The Norberg angle (NA) plays a crucial role in evaluating hip joint conformation, particularly in canines, by quantifying femoral head subluxation within the hip joint. Therefore, it is an important metric for evaluating hip joint quality and diagnosing canine hip dysplasia, the most prevalent hereditary orthopedic disorder in dogs. While contemporary tools offer automated quantification of the NA, their usage typically entails manual labeling and verification of radiographic images by professional veterinarians. To enhance efficiency and streamline this process, the study aims to develop a tool capable of predicting the NA directly from the image without the need for veterinary intervention. Due to the challenges in acquiring annotated, diverse, high-quality images, this study introduces diffusion models to expand the training dataset from 219 to 1493 images, encompassing original images. This augmentation enhances the dataset's diversity and scale, thereby improving the accuracy of Norberg angle estimation. The model predicts four key points: the center of left and right femoral heads and the edge of the left and right acetabulum, as well as the radii of the femoral heads and the Norberg angles. By evaluating 18 distinct pretrained ImageNet models, we investigate their performance pre- and post-incorporating augmented data from generated images. The results demonstrate a significant enhancement, with an average 35.3% improvement based on mean absolute percentage error when utilizing generated images from diffusion models. This study showcases the potential of diffusion modeling in advancing canine hip dysplasia diagnosis and underscores the value of incorporating augmented data to elevate model accuracy.

Prospective Evaluation of Accelerated Brain MRI Using Deep Learning-Based Reconstruction: Simultaneous Application to 2D Spin-Echo and 3D Gradient-Echo Sequences.

To prospectively evaluate the effect of accelerated deep learning-based reconstruction (Accel-DL) on improving brain magnetic resonance imaging (MRI) quality and reducing scan time compared to that in conventional MRI. This study included 150 participants (51 male; mean age 57.3 ± 16.2 years). Each group of 50 participants was scanned using one of three 3T scanners from three different vendors. Conventional and Accel-DL MRI images were obtained from each participant and compared using 2D T1- and T2-weighted and 3D gradient-echo sequences. Accel-DL acquisition was achieved using optimized scan parameters to reduce the scan time, with the acquired images reconstructed using U-Net-based software to transform low-quality, undersampled k-space data into high-quality images. The scan times of Accel-DL and conventional MRI methods were compared. Four neuroradiologists assessed the overall image quality, structural delineation, and artifacts using Likert scale (5- and 3-point scales). Inter-reader agreement was assessed using Fleiss' kappa coefficient. Signal-to-noise ratio (SNR) and contrast-to-noise ratio (CNR) were calculated, and volumetric quantification of regional structures and white matter hyperintensities (WMHs) was performed. Accel-DL showed a mean scan time reduction of 39.4% (range, 24.2%-51.3%). Accel-DL improved overall image quality (3.78 ± 0.71 vs. 3.36 ± 0.61, P < 0.001), structure delineation (2.47 ± 0.61 vs. 2.35 ± 0.62, P < 0.001), and artifacts (3.73 ± 0.72 vs. 3.71 ± 0.69, P = 0.016). Inter-reader agreement was fair to substantial (κ = 0.34-0.50). SNR and CNR increased in Accel-DL (82.0 ± 23.1 vs. 31.4 ± 10.8, P = 0.02; 12.4 ± 4.1 vs. 4.4 ± 11.2, P = 0.02). Bland-Altman plots revealed no significant differences in the volumetric measurements of 98.2% of the relevant regions, except in the deep gray matter, including the thalamus. Five of the six lesion categories showed no significant differences in WMH segmentation, except for leukocortical lesions (r = 0.64 ± 0.29). Accel-DL substantially reduced the scan time and improved the quality of brain MRI in both spin-echo and gradient-echo sequences without compromising volumetry, including lesion quantification.

WGAN-GP for Synthetic Retinal Image Generation: Enhancing Sensor-Based Medical Imaging for Classification Models.

Accurate synthetic image generation is crucial for addressing data scarcity challenges in medical image classification tasks, particularly in sensor-derived medical imaging. In this work, we propose a novel method using a Wasserstein Generative Adversarial Network with Gradient Penalty (WGAN-GP) and nearest-neighbor interpolation to generate high-quality synthetic images for diabetic retinopathy classification. Our approach enhances training datasets by generating realistic retinal images that retain critical pathological features. We evaluated the method across multiple retinal image datasets, including Retinal-Lesions, Fine-Grained Annotated Diabetic Retinopathy (FGADR), Indian Diabetic Retinopathy Image Dataset (IDRiD), and the Kaggle Diabetic Retinopathy dataset. The proposed method outperformed traditional generative models, such as conditional GANs and PathoGAN, achieving the best performance on key metrics: a Fréchet Inception Distance (FID) of 15.21, a Mean Squared Error (MSE) of 0.002025, and a Structural Similarity Index (SSIM) of 0.89 in the Kaggle dataset. Additionally, expert evaluations revealed that only 56.66% of synthetic images could be distinguished from real ones, demonstrating the high fidelity and clinical relevance of the generated data. These results highlight the effectiveness of our approach in improving medical image classification by generating realistic and diverse synthetic datasets.

Advancing Image Synthesis: Methods and Applications of Latent Diffusion Models

Diffusion models (DMs) have revolutionized the field of generative modelling, delivering exceptional results in tasks such as image synthesis, inpainting, and super-resolution. Despite their success, the reliance on pixel-space processing in these models has imposed substantial computational challenges, requiring hundreds of GPU days for training and significant resources for inference. In this work, we introduce Latent Diffusion Models (LDMs), an innovative framework that addresses these limitations by operating within a perceptually compressed latent space derived from a pretrained autoencoder. This paradigm shift significantly reduces the computational complexity of both training and inference while maintaining the high fidelity and diversity of the generated outputs. LDMs leverage a two-stage approach: a pretrained autoencoder for efficient latent-space representation and a diffusion model trained directly within this space. By introducing cross-attention layers into the architecture, LDMs also support flexible conditioning on various modalities such as text descriptions, semantic maps, or lowresolution images. This versatility enables the model to perform a range of tasks, including text-to-image generation, classconditional image synthesis, and high-resolution super-resolution. Our experiments demonstrate that LDMs achieve competitive or state-of-the-art performance across multiple benchmarks, including CelebA-HQ, ImageNet, and MS-COCO, while requiring significantly fewer computational resources than pixel-space diffusion models. For instance, LDMs reduce training time by up to 2.7× and inference memory requirements by 50%, all while improving sample quality. This work highlights the potential of latent-space generative models to democratize access to advanced generative AI technologies, making them feasible for researchers and practitioners with limited computational resources. At the same time, we discuss the ethical considerations of using such models, including their potential misuse for creating manipulated content. Latent Diffusion Models pave the way for efficient, scalable, and high-quality image synthesis, providing a robust foundation for future advancements in generative modelling.

An MRI-guided stereotactic neurosurgical robotic system for semi-enclosed head coils.

Magnetic resonance imaging (MRI) offers high-quality soft tissue imaging without radiation exposure, which allows stereotactic techniques to significantly improve outcomes in cranial surgeries, particularly in deep brain stimulation (DBS) procedures. However, conventional stereotactic neurosurgeries often rely on mechanical stereotactic head frames and preoperative imaging, leading to suboptimal results due to the invisibility and the contact with patient's head, which may cause additional harm. This paper presents a frameless, MRI-guided stereotactic neurosurgical robotic system. The robot features a seven-degree-of-freedom (7-DOF) remote center of motion, with five DOFs for preoperative trajectory alignment to the target lesion and two DOFs for defining the depth and twisting motion of the needle during insertion, thus to minimize tissue damage. The system employs interactive MRI guidance for real-time visualization of the puncture process, showing great potential in reducing surgery time, enhancing targeting accuracy, and improving safety. Experiments were conducted on the proposed system to evaluate signal-to-noise ratio (SNR) and geometric distortion. During the simultaneous operation and imaging, the system demonstrated less than 10.02% SNR attenuation and less than 0.1% geometric distortion, ensuring image usability. The free-space positioning accuracy of the system was evaluated using a laser tracker, revealing a tip position repeatability error within 0.3 ± 0.1mm.

Continuous implicit neural representation for arbitrary super-resolution of system matrix in magnetic particle imaging

Abstract Objective. Magnetic particle imaging (MPI) is a novel imaging technique that uses magnetic fields to detect tracer materials consisting of magnetic nanoparticles. System Matrix (SM) based image reconstruction is essential for achieving high image quality in MPI. However, the time-consuming SM calibrations need to be repeated whenever the magnetic field's or nanoparticle's characteristics change. Accelerating this calibration process is therefore crucial. The most common acceleration approach involves undersampling during the SM calibration procedure, followed by super-resolution methods to recover the high-resolution (HR) SM. However, these methods typically require separate training of multiple models for different undersampling ratios, leading to increased storage and training time costs.&amp;#xD;Approach. We propose an arbitrary-scale SM super-resolution method based on continuous Implicit Neural Representation (INR). Using INR, the SM is modeled as a continuous function in space, enabling arbitrary-scale super-resolution by sampling the function at different densities. A cross-frequency encoder is implemented to share SM frequency information and analyze contextual relationships, resulting in a more intelligent and efficient sampling strategy. Convolutional Neural Networks (CNNs) are utilized to learn and optimize the grid sampling process in INR, leveraging the advantage of CNNs in learning local feature associations and considering surrounding information comprehensively.&amp;#xD;Main results. Experimental results on OpenMPI demonstrate that our method outperforms existing methods and enables calibration at any scale with a single model. The proposed method achieves high accuracy and efficiency in SM recovery, even at high undersampling rates.&amp;#xD;Significance. The proposed method significantly reduces the storage and training time costs associated with SM calibration, making it more practical for real-world applications. By enabling arbitrary-scale super-resolution with a single model, our approach enhances the flexibility and efficiency of MPI systems, paving the way for more widespread adoption of MPI technology.