Narrow Results

Medical imaging is considered the first step in the examination for diagnosing kidney disease. However, the recent increase in kidney stone cases has put a remarkable burden on the whole medical diagnosis system. This problem causes a demand of an automatic kidney stone detection model to decrease the weight for medical imaging phase. Moreover, X-ray imaging, which is the current most popular medical imaging technique, is struggling with false positives, which is caused by low quality, making the diagnosis more challenging. Due to the two main problems above, this study is carried out with two contributions. First, fusing different attention modules to YOLOv7 architecture that shall bring higher performance of kidney stone detection task. Second, proposing the use of super resolution (SR) models that shall address the problem of low quality in X-ray image. As a result, the proposed YOLOv7 with attention modules easily outperforms the YOLOv7 baseline in detection performance, the highest accuracy model belongs to convolution block attention module attached with YOLOv7, which reaches 91.2% mAP50. When SR models are applied to upsample X-ray images, these SR X-ray images enable the proposed attention-based models to improve the precision and sensitivity considerably, with the highest precision reaching 97.3% and highest sensitivity hitting a peak at 91.7%. Consequently, our methods are proposed to address current issues of kidney stone diagnosis and contribute another aspect of X-ray image enhancement.

Recently, many studies have shown that deep convolutional neural network can achieve superior performance in image super resolution (SR). The majority of current CNN-based SR methods tend to use deeper architecture to get excellent performance. However, with the growing depth and width of network, the hierarchical features from low-resolution (LR) images cannot be exploited effectively. On the other hand, most models lack the ability of discriminating different types of information and treating them equally, which results in limiting the representational capacity of the models. In this study, we propose the multi-attention residual network (MARN) to address these problems. Specifically, we propose a new multi-attention residual block (MARB), which is composed of attention mechanism and multi-scale residual network. At the beginning of each residual block, the channel importance of image features is adaptively recalibrated by attention mechanism. Then, we utilize convolutional kernels of different sizes to adaptively extract the multi-attention features on different scales. At the end of blocks, local multi-attention features fusion is applied to get more effective hierarchical features. After obtaining the outputs of each MARB, global hierarchical feature fusion jointly fuses all hierarchical features for reconstructing images. Our extensive experiments show that our model outperforms most of the state-of-the-art methods.

Existing super-resolution methods convert high-resolution images into low-resolution images, and use the synthesized images as input to train the model. However, it is difficult for synthetic low-resolution images to reflect the characteristics of real low-resolution images, resulting in poor model performance in practical applications. To address this problem, we propose a recurrent super-resolution framework, which consists of a degradation model and a reconstruction model. The degradation model degenerates the real high-resolution image into a more real low-resolution image, which is used as the input of the super-resolution reconstruction network, and then uses the reconstruction model to reconstruct the low-resolution image, and calculates the error with the original image. The generated high-resolution image is input into the degradation model again for degradation processing, forming a symmetrical and cyclic network structure, so that the super-resolution model has a better effect when reconstructing the real low-scoring image. In addition, the spatial attention mechanism is introduced into the generator network, which expands the receptive field of the convolution kernel, better extracts long-distance image features and improves the texture details of super-resolution images, which is consistent with the global.

In this paper, an effort is made to propose an effective image super resolution (SR) approach to recover a high resolution (HR) image from a single low resolution (LR) image. This approach is based on an iterative back projection (IBP) method with the edge preserving infinite symmetric exponential filter (ISEF) and difference image. Amalgamation of ISEF and difference image provides high frequency information. This approach is applied on different type of images and compared results with different existing image SR approaches. Simulation results demonstrate that proposed approach can more precisely enlarge the LR image. This proposed approach decreases mean square error (MSE) and mean absolute error (MAE) and increases the peak signal-to-noise ratio (PSNR) significantly compared to other existing approaches.

Downsampling performed prior to encoding in video compression discards the high-frequency information of the original video frames. This necessitates super resolution (SR) to recover the lost information after decoding. In this paper, we propose a framework for the downsampling-based video compression problem along with a method for wavelet-based SR. We develop a framework by integrating the downsampling and transform steps of the conventional models. We then use the proposed SR method implemented directly on the wavelet sub-bands at the decoder side. Experimental results demonstrate the superiority of the proposed method in comparison with the conventional techniques employed for resolution enhancement of decoded videos.

In this paper, we propose a method to realize single image super resolution using a network composed of residual blocks with a cross architecture and wavelet transform. For the high quality of the reconstructed high resolution image, the brightness change must be sharp at the edge and gentle at the flat region. In single image super resolution, it is important to accurately predict the high-frequency components of the image. Hence, we propose a method to realize single image super resolution by estimating the high-frequency component in the wavelet domain using neural network composed of residual blocks with two crossing skip connections.

Fluorescence polarization is related to the dipole orientation of chromophores, making fluorescence polarization microscopy possible to reveal structures and functions of tagged cellular organelles and biological macromolecules. Several recent super resolution techniques have been applied to fluorescence polarization microscopy, achieving dipole measurement at nanoscale. In this review, we summarize both diffraction limited and super resolution fluorescence polarization microscopy techniques, as well as their applications in biological imaging.

This paper presents the development of a magnetic resonance imaging (MRI)-conditional needle positioning robot designed for spinal cellular injection. High-accuracy targeting performance is achieved by the combination of a high precision, parallel-plane, needle-orientation mechanism utilizing linear piezoelectric actuators with an iterative super-resolution (SR) visual navigation algorithm using multi-planar MR imaging. In previous work, the authors have developed an MRI conditional robot with positioning performance exceeding the standard resolution of MRI, rendering the MRI resolution the limit for navigation. This paper further explores the application of SR to images for robot guidance, evaluating positioning performance through simulations and experimentally in benchtop and MRI experiments.