Design of Network Cascade Structure for Image Super-Resolution

Image super resolution is an important field of computer research.The current mainstream image super-resolution technology is to use deep learning to mine the deeper features of the image,and then use it for image restoration.However,most of these models mentioned above only trained the images in a specific scale and do not consider the relationships between different scales of images.In order to utilize the information of images at different scales,we design a cascade network structure and cascaded super-resolution convolutional neural networks.This network contains three cascaded FSRCNNs.Due to each sub FSRCNN can process a specific scale image,our network can simultaneously exploit three scale images,and can also use the information of three different scales of images.Experiments on multiple datasets confirmed that the proposed network can achieve better performance for image SR.

Determination of the available energy,standardized ileal digestibility of amino acids of fermented corn germ meal replacing soybean meal in growing pig diets

Three experiments were conducted to compare the digestible(DE),metabolizable energy(ME),and standardized ileal digestibility(SID)of amino acids(AA)in fermented corn germ meal(FCGM)and soybean meal(SBM),and evaluate the effects of FCGM replacing SBM in growing pig diets.In Exp.1,18 barrows with initial body weight(BW)of 60.2±3.40 kg were randomly allotted to 3 treatments with 6 replicates per treatment.The control diet used corn as the only energy ingredient,and 2 test diets were made by replacing 25.8%of corn with FCGM or SBM.The DE and ME of FCGM were less(P<0.01)than those of SBM.In Exp.2,18 barrows(59.3±2.52 kg BW)with ileal T-cannulas were randomly allotted to 3 treatments with 6 replicates per treatment.The SID of arginine,tryptophan and proline were greater(P<0.05)and the tyrosine was less(P=0.01)in SBM compared with FCGM.In Exp.3,144 growing pigs(56.1±5.22 kg BW)were randomly allotted to 4 treatments with 6 pens per treatment(3 barrows and 3 gilts per pen).Four diets(FCGM0,FCGM10,FCGM20 and FCGM30)were formulated using FCGM to replace 0%,10%,20%and 30%of SBM,respectively.The ME and SID values of AA of SBM and FCGM were determined by Exp.1 and 2.Results showed that increasing FCGM inclusion quadratically(P<0.05)increased the average daily gain(ADG),average daily feed intake,and the levels of serum immunoglobulin G(IgG)and urea nitrogen,and linearly(P<0.05)increased the serum IgM,the propanoic acid,butyric acid,total volatile fatty acid(VFA)and the Shannon index of microbiota in feces.Besides,the relative abundance of genus Streptococcus in FCGM0,Lactobacillus in FCGM10 and Lachnospiraceae in FCGM30 were increased(P<0.05)compared with other treatments.In conclusion,we recommend replacing 11.80%of SBM with FCGM to obtain the optimal ADG of growing pigs.Moreover,as the ratio of FCGM replacing SBM increased in diet,the immunity,intestinal microbiota and total VFA composition of growing pigs were improved.