In view of the problems of current tile defect detection mainly relying on manual detection， such as strong subjectivity， low efficiency， and high labor intensity， an improved lightweight algorithm for detecting small defects in large-format ceramic tile images based on YOLOv8 was proposed. Firstly， the high-resolution large-format image was cropped， and HorBlock was introduced into the backbone network to enhance model’s capture capability. Secondly， Large Separable Kernel Attention （LSKA） was incorporated to improve C2f for improving the detection performance of the model and model’s feature extraction capability was enhanced by introducing SA （Shuffle Attention）. Finally， Omni-Dimensional Dynamic Convolution （ODConv） was introduced to further enhance model’s capability to handle with small defects. Experimental results on Alibaba Tianchi tile defect detection dataset show that the improved model not only has lower parameters than the original YOLOv8n， but also has an increase of 8.2 percentage points in mAP@0.5 and an increase of 7 percentage points in F1 score compared to the original YOLOv8n. It can be seen that the improved model can identify and process small surface defects of large-format tiles more accurately， and improve the detection effect significantly while maintaining lightweight.

Deep learning based algorithms such as YOLO （You Only Look Once） and Faster Region-Convolutional Neural Network （Faster R-CNN） require a huge amount of training data to ensure the precision of the model， and it is difficult to obtain data and the cost of labeling data is high in many scenarios. And due to the lack of massive training data， the detection range is limited. Aiming at the above problems， a few-shot object Detection algorithm based on Siamese Network was proposed， namely SiamDet， with the purpose of training an object detection model with certain generalization ability by using a few annotated images. Firstly， a Siamese network based on depthwise separable convolution was proposed， and a feature extraction network ResNet-DW was designed to solve the overfitting problem caused by insufficient samples. Secondly， an object detection algorithm SiamDet was proposed based on Siamese network， and based on ResNet-DW， Region Proposal Network （RPN） was introduced to locate the interested objects. Thirdly， binary cross entropy loss was introduced for training， and contrast training strategy was used to increase the distinction among categories. Experimental results show that SiamDet has good object detection ability for few-shot objects， and SiamDet improves AP₅₀ by 4.1% on MS-COCO 20-way 2-shot and 2.6% on PASCAL VOC 5-way 5-shot compared with the suboptimal algorithm DeFRCN （Decoupled Faster R-CNN）.

Aiming at the problem of the jump existed in the first frame of human motion synthesis method based on Recurrent Neural Network （RNN）， which affects the quality of generated motion， a human motion synthesis method with hidden state initialization was proposed. The initial hidden state was used as independent variable， the objective function of the neural network was used as optimization goal， and the gradient descent method was used to optimize and solve the problem to obtain a suitable initial hidden state. Compared with Encoder-Recurrent-Decoder （ERD） model and Residual Gate Recurrent Unit （RGRU） model， the proposed method with initial hidden state estimation reduces the prediction error of the first frame by 63.51% and 6.90% respectively， and decreases the total error of 10 frames by 50.00% and 4.89% respectively. Experimental results show that the proposed method is better than the method without initial hidden state estimation in both motion synthesis quality and motion prediction accuracy. And the proposed method accurately estimates the hidden state of the first frame of RNN-based human motion model， which improves the quality of motion synthesis and provides reliable data support for action recognition model in real-time security monitoring.

Genotype imputation can compensate for the missing due to technical limitations by estimating the sample regions that are not covered in gene sequencing data with imputation， but the existing deep learning-based imputation methods cannot effectively capture the linkage among complete sequence loci， resulting in low overall imputation accuracy and high dispersion of batch sequence imputation accuracy. Therefore， FCSA （Fusing Convolution and Self-Attention）， an imputation method that fuses convolution and self-attention mechanism， was proposed to address the above problems， and two fusion modules were used to form encoder and decoder to construct network model. In the encoder fusion module， a self-attention layer was used to obtain the correlation among complete sequence loci， and the local features were extracted through the convolutional layer after fusing the correlation to global loci. In the decoder fusion module， the local features of the encoded low-dimensional vector were reconstructed by convolution， and the complete sequence was modeled and fused by self-attention layer. The genetic data of multiple species of animals were used for model training， and the comparison and validation were carried out on Dog， Pig and Chicken datasets. The results show that compared to SCDA （Sparse Convolutional Denoising Autoencoders）， AGIC （Autoencoder Genome Imputation and Compression） and U-net， FCSA achieves the highest average imputation accuracy at 10%， 20% and 30% missing rate. Ablation experimental results also show that the design of the two fusion modules is effective in improving the accuracy of genotype imputation.

In view of the problems of vascular pleomorphism on transverse sections and sampling imbalance in the process of detection， an improved Libra Region-Convolutional Neural Network （R-CNN） cerebral arterial stenosis detection algorithm was proposed to detect internal carotid artery and vertebral artery stenosis in Computed Tomography Angiography （CTA） images. Firstly， ResNet50 was used as the backbone network in Libra R-CNN， Deformable Convolutional Network （DCN） was introduced into the 3， 4， 5 stages of backbone network， and the offsets were learnt to extract the morphological features of blood vessels on different transverse sections. Secondly， the feature maps extracted from the backbone network were input into Balanced Feature Pyramid （BFP） with the Non-local Neural Network （Non-local NN） introduced for deeper feature fusion. Finally， the fused feature maps were input to the cascade detector， and the final detection result was optimized by increasing the Intersection-over-Union （IoU） threshold. Experimental results show that compared with Libra R-CNN algorithm， the improved Libra R-CNN detection algorithm increases 4.3， 1.3， 6.9 and 4.0 percentage points respectively in AP， AP₅₀， AP₇₅ and AP_S， respectively_{on the cerebral artery CTA dataset； on the public CT dataset of colon polyps， the improved Libra R-CNN detection algorithm has the AP， AP50， AP75 and APS increased by 6.6， 3.6， 13.0 and 6.4 percentage points， respectively. By adding DCN， Non-local NN and cascade detector to the backbone network of Libra R-CNN algorithm， the features are further fused to learn the semantic information of cerebral artery structure and make the results of narrow area detection more accurate， and the improved algorithm has the ability of generalization in different detection tasks.}

At present, the most common way for bitcoin mining is miners joining in a pool. However, there is a phenomenon that the mining pools penetrate each other, which will result in a decrease in the miners' income of the attacked pools, and a reduction in computing power of the attacking pools. Therefore, the overall computing power of the bitcoin system is reduced. Aiming at the problem of mutual attack and non-cooperative mining between mining pools, an Adaptive Zero-Determinant strategy (AZD) was proposed to promote the cooperation of miners. The strategy adopted the idea of comparing expected payoff with cooperation and defection in the next round then choosing a strategy with high payoff. Firstly, miners' payoff in the next round under two situations could be predicted by the combination of Temporal Difference Learning Method (TD(λ)) and Zero-Determinant strategy (ZD). Secondly, by comparing the cooperation payoff with defection payoff in the next round, a more favorable strategy was chosen for miners by Decision Making Process (DMP), so the cooperation probability and defection probability in the next round were changed correspondingly. Finally, through the iterative implementation of AZD strategy, the ming pools in the network would cooperate with each other and mine actively. Simulation results show that compared with adaptive strategy, AZD strategy increases the speed of converging cooperation probability to 1 by 36.54%, compared with ZD strategy, it improves the stability by 50%. This result indicates that AZD strategy can effectively promote the cooperation of miners, improve the convergence rate of cooperation and ensure the stable income of mining pools.

Concerning the problems that coutour complex path trajectory generation and low slicing efficiency in rapid prototyping of artificial bone tissue, a method to simplify the slicing process of triangle mesh was proposed in this paper. The medical image sequences were reconstructed by the Marching Cubes (MC) algorithm, the triangle meshes were grouped into triangle arrays according to the order of the reconstruction process. And then, the intersection points between the slice plane and the triangle array were calculated by edge tracking. It was found that the slicing efficiency of simplified process was increased by 4.65% on average compared with the triangular mesh STereoLithography (STL) model. The experimental results indicate that the proposed method can generate contour data used for 3D printing directly from medical image sequences of human bone tissu, so as to realize the rapid prototyping of bone tissue.

In order to improve the processing ability for uncertainty data using the traditional Fuzzy Support Vector Machine (FSVM), FSVM with fuzzy similarity measure and high dimensional space fuzzy mapping was proposed. Firstly, by using Gregson similarity measure, the fuzzy similarity measure function was established, which was effective to explain the uncertainty information. And then, using the theory of mapping and Mercer, fuzzy similarity kernel learning was formulated and used in the algorithm of the FSVM. Finally, this algorithm was used to the modeling of the material removal rate in the rotary ultrasonic machining with uncertainty data. Compared to the results using traditional FSVM methods, the current approach can better process uncertainty data with less operation steps. And the proposed method has higher accuracy in processing uncertainty data with lower computational complexity.

The trust management models in Peer-to-Peer (P2P) network mainly have two problems. For one thing, the different influences on value of trust between short-term trading and long-term trading are usually ignored. For another, the lack of the specific risk analysis on trading resources exists. Consequently, focusing on the quality of different nodes and its opposite risk value, this paper introduced the concept of risk factor with setting up its value and proposed a trust management model based on evaluation of value-at-risk with changing time. From the simulation results, a higher efficiency on resisting malicious actions in P2P network is achieved, and it has confirmed to select better traders effectively with a deeply quantitative analysis of trade resources through the model.

Concerning the problem of Track-Before-Detect (TBD) in a multi-target environment, in this paper, a TBD algorithm based on Joint Multi-target Probability Density (JMPD) model was proposed. The JMPD was a single probabilistic entity that captured uncertainty about the number of targets present in the surveillance region as well as their individual states and a Particle Filter (PF) was used to recursively estimate the JMPD. The simulation results demonstrate that the birth and death of target can be estimated accurately as well as its trajectory by the proposed algorithm with smaller detection delays.

In order to improve the accuracy and stability of the players in the movement of RoboCup3D simulation platform, a kind of real-time resolution of inverse kinematics for humanoid NAO model was proposed. Firstly, the lower limb topology of NAO model was analyzed, and its forward kinematics model was established. Secondly, the equations of every joint angle in all lower limbs were derived by real-time inverse kinematics analytic method. Finally, the algorithm was realized by coding. The experimental results validate the numerical stability and feasibility of online execution of the method, and the overall competitive level of the RoboCup3D simulation soccer team has been enhanced.

Since the Particle Filter-based Track-Before-Detect (PF TBD) is subject to severe sample impoverishment, the fission reproduction PF TBD algorithm was proposed. To incorporate TBD problem, the particles were divided into three types according to an existence variable which indicates the presence/absence of a target in the data. Three types of particles were death, birth and survival, respectively, and the survival particles were processed by the fission reproduction. The process increases the diversity of particles, and overcomes sample impoverishment. The simulation results demonstrate that, compared to the PF TBD, the proposed algorithm can provide stable and reliable detection as well as accurate tracking.