Compound feature attention network with edge enhancement for low-dose CT denoising.

BACKGROUND: Low-dose CT (LDCT) images usually contain serious noise and artifacts, which weaken the readability of the image. OBJECTIVE: To solve this problem, we propose a compound feature attention network with edge enhancement for LDCT denoising (CFAN-Net), which consists of an edge-enhanced module and a proposed compound feature attention block (CFAB). METHODS: The edge enhancement module extracts edge details with the trainable Sobel convolution. CFAB consists of an interactive feature learning module (IFLM), a multi-scale feature fusion module (MFFM), and a joint attention module (JAB), which removes noise from LDCT images in a coarse-to-fine manner. First, in IFLM, the noise is initially removed by cross-latitude interactive judgment learning. Second, in MFFM, multi-scale and pixel attention are integrated to explore fine noise removal. Finally, in JAB, we focus on key information, extract useful features, and improve the efficiency of network learning. To construct a high-quality image, we repeat the above operation by cascading CFAB. RESULTS: By applying CFAN-Net to process the 2016 NIH AAPM-Mayo LDCT challenge test dataset, experiments show that the peak signal-to-noise ratio value is 33.9692 and the structural similarity value is 0.9198. CONCLUSIONS: Compared with several existing LDCT denoising algorithms, CFAN-Net effectively preserves the texture of CT images while removing noise and artifacts.

Transfer of CeO2 nanoparticles between freshwater omnivorous organisms: Effect of feces and necrophagy.

Transfer of CeO2 engineered nanoparticles (NPs) through feces was investigated between two omnivorous organisms, red crucian carp (Carassius auratus red var.) and crayfish (Procambarus clarkii). Upon water exposure (5 mg/L, 7 days), the highest bioaccumulation was observed in carp gills (5.95 µg Ce/g D.W.) and crayfish hepatopancreas (648 µg Ce/g D.W.), with the bioconcentration factors (BCFs) at 0.45 and 3.61, respectively. In addition, 97.4% and 73.0% of ingested Ce were excreted by carp and crayfish, respectively. The feces of carp and crayfish were collected and fed to crayfish and carp, respectively. After feces exposure, bioconcentration was observed in both carp (BCF, 3.00) and crayfish (BCF, 4.56). After feeding crayfish with carp bodies (1.85 µg Ce/g D.W.), CeO2 NPs were not biomagnified (biomagnification factor, 0.28). Upon water exposure, CeO2 NPs were transformed into Ce(III) in the feces of both carp (24.6%) and crayfish (13.6%), and the transformation was stronger after subsequent feces exposure (100% and 73.7%, respectively). Feces exposure lowered histopathological damage, oxidative stress, and nutritional quality (e.g., crude proteins, microelements, amino acids) to carp and crayfish in comparison with water exposure. This research highlights the importance of feces exposure on the transfer and fate of NPs in aquatic ecosystems.

Learnable PM diffusion coefficients and reformative coordinate attention network for low dose CT denoising.

Objective.Various deep learning methods have recently been used for low dose CT (LDCT) denoising. Aggressive denoising may destroy the edge and fine anatomical structures of CT images. Therefore a key issue in LDCT denoising tasks is the difficulty of balancing noise/artifact suppression and edge/structure preservation.Approach.We proposed an LDCT denoising network based on the encoder-decoder structure, namely the Learnable PM diffusion coefficient and efficient attention network (PMA-Net). First, using the powerful feature modeling capability of partial differential equations, we constructed a multiple learnable edge module to generate precise edge information, incorporating the anisotropic image processing idea of Perona-Malik (PM) model into the neural network. Second, a multiscale reformative coordinate attention module was designed to extract multiscale information. Non-overlapping dilated convolution capturing abundant contextual content was combined with coordinate attention which could embed the spatial location information of important features into the channel attention map. Finally, we imposed additional constraints on the edge information using edge-enhanced multiscale perceptual loss to avoid structure loss and over-smoothing.Main results.Experiments are conducted on simulated and real datasets. The quantitative and qualitative results show that the proposed method has better performance in suppressing noise/artifacts and preserving edges/structures.Significance.This work proposes a novel edge feature extraction method that unfolds partial differential equation into neural networks, which contributes to the interpretability and clinical application value of neural network.

Intelligent anti-epidemic mask based on KF and ECF fusion algorithm.

In response to environmental pollution and the spread of Coronavirus Disease 2019 (COVID-19), this paper proposes a new type of smart mask design, and specifically proposes an optimized double closed-loop control method, especially an improved filtering fusion algorithm. Using the filtering fusion algorithm proposed in this paper, after the Kalman filter (KF) filters the raw data of the attitude sensor, explicit complementary filtering and data fusion are used to obtain the attitude angle of the body. At the same time, the obtained attitude angle is combined with acceleration and blood oxygen concentration to obtain the behaviour characteristic value. On this basis, the speed of the oxygen supply fan captured by the photoelectric sensor is used to form a closed loop with the characteristic value of the behaviour. Finally, the structure of the mask is upgraded and optimized through fluid mechanics simulation, and experiments have verified that the combination of the replaceable filter cloth, the intelligent control system and the ultraviolet disinfection device can effectively protect people's health.

[The NIR spectra based variety discrimination for single soybean seed].

With the development of soybean producing and processing, the quality breeding becomes more and more important for soybean breeders. Traditional sampling detection methods for soybean quality need to destroy the seed, and does not satisfy the requirement of earlier generation materials sieving for breeding. Near infrared (NIR) spectroscopy has been widely used for soybean quality detection. However, all these applications were referred to mass samples, and they were not suitable for little or single seed detection in breeding procedure. In the present study, the acousto--optic tunable filter (AOTF) NIR spectroscopy was used to measure the single soybean seed. Two varieties of soybean were measured, which contained 60 KENJIANDOU43 seeds and 60 ZHONGHUANG13 seeds. The results showed that NIR spectra combined with soft independent modeling of class analogy (SIMCA) could accurately discriminate the soybean varieties. The classification accuracy for KENJIANDOU43 seeds and ZHONGHUANG13 was 100%. The spectra of single soybean seed were measured at different positions, and it showed that the seed shape has significant influence on the measurement of spectra, therefore, the key point for single seed measurement was how to accurately acquire the spectra and keep their representativeness. The spectra for soybeans with glossy surface had high repeatability, while the spectra of seeds with external defects had significant difference for several measurements. For the fast sieving of earlier generation materials in breeding, one could firstly eliminate the seeds with external defects, then apply NIR spectra for internal quality detection, and in this way the influence of seed shape and external defects could be reduced.

[The preliminary study for testing freshness of meat by using near-infrared reflectance spectroscopy].

The value of the volatile basic nitrogen of meat is an important index to determine the freshness of meat. It is difficult to meet the demand of fast and non-destructive measurement by means of classical semimicro-quantitative nitrogent method. A model to predict the value of the volatile basic nitrogen based on near-infrared reflectance spectroscopy was established. Cluster analysis methods were applied to deal with the data of NIRS. If the content of TVB-N is more than 11. 6 mg x (100 g)(-1), the back pork may be rotten. The result shows that using NIRS could indicate the freshness of meat quickly and non-invasively.