Image Quality Degradation Research Articles

RCFormer: radiation correction method for degraded multispectral UAV images using vision transformers

ABSTRACT In modern remote sensing technology, Unmanned Aerial Vehicles (UAVs) have become crucial data acquisition platforms. However, low-altitude remote sensing images from UAVs are often affected by atmospheric scattering and absorption, particularly under conditions like haze, which severely degrade image quality. Traditional radiation correction methods cannot accurately capture the true reflectance of ground objects under these conditions. This paper introduces RCFormer, an innovative Transformer-based network architecture. It combines the self-attention mechanism of Swin Transformer with a novel network structural design, especially incorporating parallel convolutional structures to enhance the extraction capability of surface features. Additionally, with the integration of atmospheric characteristic refinement module (ACRM) and Multi-scale Feature Fusion Gate (MFFG), the proposed RCFormer not only captures atmospheric information from deep features but also effectively fuses features across different scales, thereby improving the accuracy, transferability and robustness of the radiation correction model. Finally, quantitative and qualitative evaluations using real-world data demonstrate that RCFormer outperforms other dehazing networks in both degraded and haze-free images while also offering good efficiency and ease of operation.

A whole gamma imaging prototype for higher quantitative imaging of 89Zr-labeled antibodies in a tumor mouse model

Objective.Positron emission tomography (PET) has become an important clinical modality, but it is limited to imaging the annihilation radiation from positron-electron collisions. Recently, PET imaging with89Zr, which has a half-life of 3 d, has attracted much attention in immuno-PET to visualize immune cells and cancer cells by targeting specific antibodies on the cell surface. However,89Zr emits a single gamma ray at 909 keV four times more frequently than positrons, causing image quality (IQ) degradation in conventional PET. To overcome this drawback, use of such single gamma rays for imaging was previously proposed as whole gamma imaging (WGI). In WGI, a single gamma ray is detected by Compton imaging; by inserting a scatter detector ring inside the PET ring, WGI can realize both PET imaging and Compton imaging in one modality. A prototype for WGI was developed and Compton imaging of a mouse after intravenous administration of89Zr oxalate was demonstrated. However, the Compton imaging of the single gamma ray still presented a challenge due to its low IQ compared to PET.Approach.In this study, the scatter detector insert of the earlier WGI prototype was redesigned with the aim of improving Compton imaging performance. The new prototype produced WGI images by additive averaging of PET and Compton images after optimizing the ratio of each iteration in the image reconstruction. WGI IQ was then evaluated using the NEMA NU4 IQ phantom, and a tumor-burdened mouse was imaged with WGI up to 12 d after89Zr labeled antibody injection.Main results.Consequently, the Compton imaging performance was improved by lowering the angular resolution measure from 6.7 degrees to 6.4 degrees and the sensitivity from 0.11% to 0.18% compared to the previous prototype WGI. The phantom images with WGI showed a 15% reduction in noise and a 3% increase in contrast recovery under low-statistical conditions compared to images reconstructed by PET data alone.Significance. In-vivomouse imaging with the new prototype WGI was successfully performed. This successful imaging leads to the expectation that future whole-body WGI imaging will enable more sensitive and better quantitative89Zr antigen-antibody reaction imaging to be obtained.

Remote sensing image destriping with two-stage image decomposition network

ABSTRACT Stripe noise often affects remote sensing images, leading to the degradation of imaging quality and impacting subsequent image processing. Deep learning methods have made remarkable advancements in removing stripes from remote sensing images with their powerful feature extraction capability. However, it is noteworthy that these methods are still insufficient in analysing structural characteristics in the direction of stripes and multi-scale contextual information. To deal with the above issues, we propose a remote sensing image destriping with a two-stage image decomposition network, named TSIDNet, which utilizes the stripe structural characteristics to effectively remove stripe noise while retaining better details. Firstly, stripes are directional, and the structural information of the stripes is mainly concentrated in the image after the differential decomposition of the direction along the stripes. Therefore, a differential decomposition stripe extraction subnetwork (DDSESN) is constructed to generate latent images. Within this subnetwork, we further design a multi-scale cross-fusion residual block (MCRB) and cross-scale fusion attention block (CSFAB) to gradually expand the network’s receptive field, which is conducive to extract and remove stripes more completely. Secondly, the features obtained by adding the latent clear image with details extracted from the stripe’s reverse direction are input into a DWT decomposition detail enhancement subnetwork (DDDESN), which utilizes the discrete wavelet transform (DWT) and the dense residual network for detail enhancement. Demonstrated through experiments, the proposed TSIDNet enables the removal of image stripes while preserving details, and surpassing the comparative methods in qualitative as well as quantitative evaluations. The code will be provided after acceptance.

Depth‐Modulus Engineered Meta‐Elastomers for Multiaxial Strain Programming in Stretchable Displays

With the growing demand for displays with diverse form factors, stretchable displays are at the forefront of this evolution garnering significant interest. One challenge is image warping caused by the high Poisson's ratio (ν) of conventional elastomers during stretching, meta‐elastomer (ME) substrates incorporating high‐modulus mechanical metamaterial frames with a negative ν have emerged as promising solutions. However, the modulus mismatch between the frame and matrix results in undesirable thickness variations in ME substrates under tensile stress, degrading image quality by distorting the focal length. In this study, depth‐modulus engineered MEs that achieve near‐zero ν under multiaxial strain are reported. By infiltrating a vaporized cross‐linker, a continuous depth‐modulus gradient in the ME substrates is created, effectively minimizing Poisson's effects and z‐axis deformation simultaneously. These continuous gradient depth‐modulus engineered ME (C‐GDME) substrates reduce the in‐plane ν to 0.09 and exhibit thickness variation close to half that of pristine ME substrates without modulus tailoring. Furthermore, C‐GDME substrates demonstrate excellent mechanical durability and ≈90% visible‐light transmittance. The integrated light‐emitting diode arrays on the C‐GDME substrates display linear and predictable pixel displacement under stretching, highlighting their potential for stretchable displays.

Attempt to Improve MIP Image of Cerebrovascular MRA by Subtraction Method: Efficient Acquisition of Mask Images Using Segmented TOF and Compressed Sensing

The aim of this study was to investigate imaging conditions that allow for the rapid acquisition of mask images used in the subtraction method, one of the depiction improvement methods for brain magnetic resonance angiography, by employing compressed sensing (CS) combined with segmented time-of-flight (TOF). The experiment was performed on healthy volunteers using 3.0T-MRI. We examined the blood flow signal suppression effects in arteries and veins at different numbers of segments. We determined the flip angle at which the signal value of fat in the mask image became equivalent to that in the angio image. We visually assessed the impact of mask images with varying CS factors on subtraction images and determined the optimal CS factor. Effective blood flow signal suppression in arteries and veins was achieved with segment numbers of 12 or fewer, and a flip angle of 18° resulted in equivalent fat signal values. Even with a CS factor set to its maximum value of 12, no degradation in image quality was observed in the maximum intensity projection images. Using segmented TOF with CS, optimal imaging conditions for acquiring mask images for the subtraction method were determined to be segment number 9, flip angle 18°, and CS factor 12.

Predicting Image Quality Degradation as a Result of Two-Phase Sample Transport in LA-ICP-TOFMS Mapping of Carbon-Based Materials

Recent findings reported by Van Helden et al. (Anal. Chim. Acta, 2024, 1287, 342089) have revealed that a whole suite of elements (S, Zn, As, Se, Cd, I, Te and...

Surface defect detection of steel strip at low resolution based on SAC-YOLOv5

Abstract Surface defects are common occurrences in the production process of strip steel. The development of automatic and efficient intelligent detection algorithms for strip steel is crucial for enhancing product quality and operational safety. While deep learning-based defect detection techniques have achieved satisfactory accuracy when applied to high-resolution images, obtaining a sufficient number of high-resolution images in practical engineering scenarios is challenging. The degradation of image quality often results in a significant decline in the performance of existing detection techniques. To address these challenges, this paper proposes SAC-you only look once (YOLO)v5-based surface defect detection model specifically designed for low-resolution strip steel images. SAC, SIoU based K-Means++, asymmetric convolutional neural networks (ACNNs) and composite down-sampling feature fusion block (CDFFB). Firstly, we introduce an anchor boxes clustering algorithm called Shape-IoU based K-Means++ (SIoU based K-Means++) to enhance the efficiency of anchor boxes regression. Secondly, we construct an ACNNs for multi-level feature extraction. Utilizing reparameterization techniques, we reduce the inference resources required by the model. Additionally, we propose a CDFFB which enhances key texture information and improves the model’s nonlinear fitting ability. Experimental analysis using NEU-DET surface defect data demonstrates the superiority of the proposed model in processing low-resolution strip steel surface defect images. In comparison to YOLOv8, YOLOv7, YOLOX, YOLOv3SPP, CenterNet and Faster RCNN, SAC-YOLOv5 outperforms all other models in terms of both speed and accuracy.

Hybrid Frequency–Spatial Domain Learning for Image Restoration in Under-Display Camera Systems Using Augmented Virtual Big Data Generated by the Angular Spectrum Method

In the rapidly advancing realm of mobile technology, under-display camera (UDC) systems have emerged as a promising solution for achieving seamless full-screen displays. Despite their innovative potential, UDC systems face significant challenges, including low light transmittance and pronounced diffraction effects that degrade image quality. This study aims to address these issues by examining degradation phenomena through optical simulation and employing a deep neural network model incorporating hybrid frequency–spatial domain learning. To effectively train the model, we generated a substantial synthetic dataset that virtually simulates the unique image degradation characteristics of UDC systems, utilizing the angular spectrum method for optical simulation. This approach enabled the creation of a diverse and comprehensive dataset of virtual degraded images by accurately replicating the degradation process from pristine images. The augmented virtual data were combined with actual degraded images as training data, compensating for the limitations of real data availability. Through our proposed methods, we achieved a marked improvement in image quality, with the average structural similarity index measure (SSIM) value increasing from 0.8047 to 0.9608 and the peak signal-to-noise ratio (PSNR) improving from 26.383 dB to 36.046 dB on an experimentally degraded image dataset. These results highlight the potential of our integrated optics and AI-based methodology in addressing image restoration challenges within UDC systems and advancing the quality of display technology in smartphones.

A Full-Scale Shadow Detection Network Based on Multiple Attention Mechanisms for Remote-Sensing Images

Shadows degrade image quality and complicate interpretation, underscoring the importance of accurate shadow detection for many image analysis tasks. However, due to the complex backgrounds and variable shadow characteristics of remote sensing images (RSIs), existing methods often struggle with accurately detecting shadows of various scales and misclassifying dark, non-shaded areas as shadows. To address these issues, we proposed a comprehensive shadow detection network called MAMNet. Firstly, we proposed a multi-scale spatial channel attention fusion module, which extracted multi-scale features incorporating both spatial and channel information, allowing the model to flexibly adapt to shadows of different scales. Secondly, to address the issue of false detection in non-shadow areas, we introduced a criss-cross attention module, enabling non-shadow pixels to be compared with other shadow and non-shadow pixels in the same row and column, learning similar features of pixels in the same category, which improved the classification accuracy of non-shadow pixels. Finally, to address the issue of important information from the other two modules being lost due to continuous upsampling during the decoding phase, we proposed an auxiliary branch module to assist the main branch in decision-making, ensuring that the final output retained the key information from all stages. The experimental results demonstrated that the model outperformed the current state-of-the-art RSI shadow detection method on the aerial imagery dataset for shadow detection (AISD). The model achieved an overall accuracy (OA) of 97.50%, an F1 score of 94.07%, an intersection over union (IOU) of 88.87%, a precision of 95.06%, and a BER of 4.05%, respectively. Additionally, visualization results indicated that our model could effectively detect shadows of various scales while avoiding false detection in non-shadow areas. Therefore, this model offers an efficient solution for shadow detection in aerial imagery.

Optimization of Imaging Reconnaissance Systems Using Super-Resolution: Efficiency Analysis in Interference Conditions.

Image reconnaissance systems are critical in modern applications, where the ability to accurately detect and identify objects is crucial. However, distortions in real-world operational conditions, such as motion blur, noise, and compression artifacts, often degrade image quality, affecting the performance of detection systems. This study analyzed the impact of super-resolution (SR) technology, in particular, the Real-ESRGAN model, on the performance of a detection model under disturbed conditions. The methodology involved training and evaluating the Faster R-CNN detection model with original and modified data sets. The results showed that SR significantly improved detection precision and mAP in most interference scenarios. These findings underscore SR's potential to improve imaging systems while identifying key areas for future development and further research.

Robust real-time single-pixel imaging based on a spinning mask via differential detection

Single-pixel imaging (SPI) has gained considerable attention over the past two decades but is still inadequate for imaging moving objects in practical scenarios. Recently, spinning masks have been employed for real-time SPI due to their fast modulation speeds. However, the intensity fluctuations caused by ambient light can overlap with the desired modulation from the mask, leading to a significant degradation in image quality, making it unable to function properly in environments with unstable lighting. Here we propose a rotating disc encoded with cyclic S-matrices, leveraging both reflective and transmissive signals for differential detection. Simulations and experiments demonstrate interference-resistant real-time SPI, achieving 31 fps at about 60 kHz modulation rates. This approach effectively suppresses external noise and is extendable to infrared and terahertz bands, advancing SPI’s practical applications.

Generative Adversarial Network-Based Distortion Reduction Adapted to Peak Signal-to-Noise Ratio Parameters in VVC

In order to address the issues of image quality degradation and distortion that arise in the context of video transmission coding and decoding, a method based on an enhanced version of CycleGAN is put forth. The lightweight attention module is integrated into the residual block of the generator module structure, thereby facilitating the extraction of image details and motion compensation. Furthermore, the perceptual function LPIPS loss is increased to align the image restoration effect more closely with human perception. Additionally, the network training method is modified, and the original image is divided into 128 × 128 small blocks for training, thus enhancing the network’s accuracy in restoring details. The experimental results demonstrate that the algorithm attains an average PSNR value of 30.1147 on the publicly accessible YUV sequence dataset, YUV Trace Dataset, which is a 9.02% enhancement compared to the original network. Additionally, the LPIPS value reaches 0.2639, representing a 10.42% reduction, and effectively addresses the issue of image quality deterioration during transmission.

Adaptive In-Sensor Computing for Enhanced Feature Perception and Broadband Image Restoration.

Traditional imaging systems struggle in weak or complex lighting environments due to their fixed spectral responses, resulting in spectral mismatches and degraded image quality. To address these challenges, a bioinspired adaptive broadband image sensor is developed. This innovative sensor leverages a meticulously designed type-I heterojunction alignment of 0D perovskite quantum dots (PQDs) and 2D black phosphorus (BP). This configuration enables efficient carrier injection control and advanced computing capabilities within an integrated phototransistor array. The sensor's unique responses to both visible and infrared (IR) light facilitate selective enhancement and precise feature extraction under varying lighting conditions. Furthermore, it supports real-time convolution and image restoration within a convolutional autoencoder (CAE) network, effectively countering image degradation by capturing spectral features. Remarkably, the hardware responsivity weights perform comparably to software-trained weights, achieving an image restoration accuracy of over 85%. This approach offers a robust and versatile solution for machine vision applications that demand precise and adaptive imaging in dynamic lighting environments.

Magnetic resonance image denoising for Rician noise using a novel hybrid transformer-CNN network (HTC-net) and self-supervised pretraining.

Magnetic resonance imaging (MRI) is a crucial technique for both scientific research and clinical diagnosis. However, noise generated during MR data acquisition degrades image quality, particularly in hyperpolarized (HP) gas MRI. While deep learning (DL) methods have shown promise for MR image denoising, most of them fail to adequately utilize the long-range information which is important to improve denoising performance. Furthermore, the sample size of paired noisy and noise-free MR images also limits denoising performance. To develop an effective DL method that enhances denoising performance and reduces the requirement of paired MR images by utilizing the long-range information and pretraining. In this work, a hybrid Transformer-convolutional neural network (CNN) network (HTC-net) and a self-supervised pretraining strategy are proposed, which effectively enhance the denoising performance. In HTC-net, a CNN branch is exploited to extract the local features. Then a Transformer-CNN branch with two parallel encoders is designed to capture the long-range information. Within this branch, a residual fusion block (RFB) with a residual feature processing module and a feature fusion module is proposed to aggregate features at different resolutions extracted by two parallel encoders. After that, HTC-net exploits the comprehensive features from the CNN branch and the Transformer-CNN branch to accurately predict noise-free MR images through a reconstruction module. To further enhance the performance on limited MRI datasets, a self-supervised pretraining strategy is proposed. This strategy employs self-supervised denoising to equip the HTC-net with denoising capabilities during pretraining, and then the pre-trained parameters are transferred to facilitate subsequent supervised training. Experimental results on the pulmonary HP 129Xe MRI dataset (1059 images) and IXI dataset (5000 images) all demonstrate the proposed method outperforms the state-of-the-art methods, exhibiting superior preservation of edges and structures. Quantitatively, on the pulmonary HP 129Xe MRI dataset, the proposed method outperforms the state-of-the-art methods by 0.254-0.597dB in PSNR and 0.007-0.013 in SSIM. On the IXI dataset, the proposed method outperforms the state-of-the-art methods by 0.3-0.927dB in PSNR and 0.003-0.016 in SSIM. The proposed method can effectively enhance the quality of MR images, which helps improve the diagnosis accuracy in clinical.

A foundation model for enhancing magnetic resonance images and downstream segmentation, registration and diagnostic tasks.

In structural magnetic resonance (MR) imaging, motion artefacts, low resolution, imaging noise and variability in acquisition protocols frequently degrade image quality and confound downstream analyses. Here we report a foundation model for the motion correction, resolution enhancement, denoising and harmonization of MR images. Specifically, we trained a tissue-classification neural network to predict tissue labels, which are then leveraged by a 'tissue-aware' enhancement network to generate high-quality MR images. We validated the model's effectiveness on a large and diverse dataset comprising 2,448 deliberately corrupted images and 10,963 images spanning a wide age range (from foetuses to elderly individuals) acquired using a variety of clinical scanners across 19 public datasets. The model consistently outperformed state-of-the-art algorithms in improving the quality of MR images, handling pathological brains with multiple sclerosis or gliomas, generating 7-T-like images from 3 T scans and harmonizing images acquired from different scanners. The high-quality, high-resolution and harmonized images generated by the model can be used to enhance the performance of models for tissue segmentation, registration, diagnosis and other downstream tasks.

Digital hologram denoising by filtering in the hologram plane using the Hilbert-Huang transform

Image quality degradation is one of the main problems in Digital Holography. A novel method for noise reduction based on the Hilbert-Huang transform is introduced that reduces the normalized contrast and partially removes the DC-term. It is successfully applied in the hologram plane and requires a single hologram to function. The method causes no reduction in spatial resolution. Moreover, it can be combined with existing noise-filtering methods to improve image quality even further.

Scatter correction for contrast-enhanced digital breast tomosynthesis with a dual-layer detector.

Contrast-enhanced digital breast tomosynthesis (CEDBT) highlights breast tumors with neo-angiogenesis. A recently proposed CEDBT system with a dual-layer (DL) flat-panel detector enables simultaneous acquisition of high-energy (HE) and low-energy (LE) projection images with a single exposure, which reduces acquisition time and eliminates motion artifacts. However, x-ray scatter degrades image quality and lesion detectability. We propose a practical method for accurate and robust scatter correction (SC) for DL-CEDBT. The proposed hybrid SC method combines the advantages of a two-kernel iterative convolution method and an empirical interpolation strategy, which accounts for the reduced scatter from the peripheral breast region due to thickness roll-off and the scatter contribution from the region outside the breast. Scatter point spread functions were generated using Monte Carlo simulations with different breast glandular fractions, compressed thicknesses, and projection angles. Projection images and ground truth scatter maps of anthropomorphic digital breast phantoms were simulated to evaluate the performance of the proposed SC method and three other kernel- and interpolation-based methods. The mean absolute relative error (MARE) between scatter estimates and ground truth was used as the metric for SC accuracy. DL-CEDBT shows scatter characteristics different from dual-shot, primarily due to the two energy peaks of the incident spectrum and the structure of the DL detector. Compared with the other methods investigated, the proposed hybrid SC method showed superior accuracy and robustness, with MARE of for all LE and HE projection images of different phantoms in both cranial-caudal and mediolateral-oblique views. After SC, cupping artifacts in the dual-energy image were removed, and the signal difference-to-noise ratio was improved by 82.0% for 8mm iodine objects. A practical SC method was developed, which provided accurate and robust scatter estimates to improve image quality and lesion detectability for DL-CEDBT.

Transmission Map Refinement Using Laplacian Transform on Single Image Dehazing Based on Dark Channel Prior Approach

Abstract Computer vision requires high-quality input images to facilitate image interpretation and analysis tasks. However, the image acquisition process does not always produce good-quality images. In outdoor environments, image quality is determined by weather or environmental conditions. Bad weather conditions due to pollution particles in the atmosphere such as smoke, fog, and haze can degrade image quality, such as contrast, brightness, and sharpness. This research proposes to obtain a better haze-free image from a hazy image by utilizing the Laplacian filtering and image enhancement techniques in the transmission map reconstruction based on the dark channel prior approach. Experimental results show that the proposed method could improve the visual quality of the dehazed images from 45% to 56% compared to the ground-truth images. The proposed method is also fairly competitive compared to similar methods in the same domain.

Mamba-based Thermal Image Dehazing

Atmospheric phenomena such as rain, snow, urban, forest fires, and artificial disasters can degrade image quality across various applications, including transportation, driver assistance systems, surveillance, military, and remote sensing. Image dehazing techniques aim to reduce the effects of haze, dust, fog, and other atmospheric distortions, enhancing image quality for better performance in computer vision tasks. Haze not only obscures details but also reduces contrast and color fidelity, significantly impacting the accuracy of computer vision (CV) models used in object detection, image classification, and segmentation. While thermal infrared (TIR) imaging is often favored for long-range surveillance and remote sensing due to its resilience to haze, atmospheric conditions can still degrade TIR image quality, especially in extreme environments. This paper introduces MTIE-Net, a novel Mamba-based network for enhancing thermal images degraded by atmospheric phenomena like haze and smoke. MTIE-Net leverages the Enhancement and Denoising State Space Model (EDSSM), which combines convolutional neural networks with state-space modeling for effective denoising and enhancement. We generate synthetic hazy images and employ domain-specific transformations tailored to thermal image characteristics to improve training in low-visibility conditions. Our key contributions include using the Mamba architecture with 2D Selective Scanning for thermal image enhancement, developing a specialized Enhancement and Denoising module, and creating a labeled thermal dataset simulating heavy haze. Evaluated on the M3DF dataset of long-range thermal images, MTIE-Net surpasses state-of-the-art methods in both quantitative metrics (PSNR, SSIM) and qualitative assessments of visual clarity and edge preservation. This advancement significantly improves the reliability and accuracy of critical systems used in remote sensing, surveillance, and autonomous operations by enhancing image quality in challenging environments.

Efficient Distortion-Free Neural Projector Deblurring in Dynamic Projection Mapping.

Dynamic Projection Mapping (DPM) necessitates geometric compensation of the projection image based on the position and orientation of moving objects. Additionally, the projector's shallow depth of field results in pronounced defocus blur even with minimal object movement. Achieving delay-free DPM with high image quality requires real-time implementation of geometric compensation and projector deblurring. To meet this demand, we propose a framework comprising two neural components: one for geometric compensation and another for projector deblurring. The former component warps the image by detecting the optical flow of each pixel in both the projection and captured images. The latter component performs real-time sharpening as needed. Ideally, our network's parameters should be trained on data acquired in an actual environment. However, training the network from scratch while executing DPM, which demands real-time image generation, is impractical. Therefore, the network must undergo pre-training. Unfortunately, there are no publicly available large real datasets for DPM due to the diverse image quality degradation patterns. To address this challenge, we propose a realistic synthetic data generation method that numerically models geometric distortion and defocus blur in real-world DPM. Through exhaustive experiments, we have confirmed that the model trained on the proposed dataset achieves projector deblurring in the presence of geometric distortions with a quality comparable to state-of-the-art methods.