Sperm detection performance is particularly critical for sperm motility tracking. However, there are a large number of non-sperm objects, sperm occlusion and poorly detailed texture features in semen images, which directly affect the accuracy of sperm detection. To solve the problem of false detection and missed detection in sperm detection, a multi-sperm target detection model, Yolov5s-SA, with an SA attention mechanism is proposed based on the YOLOv5s algorithm. Firstly, a depthwise, separable convolution structure is used to replace the partial convolution of the backbone network, which can ensure stable precision and reduce the number of model parameters. Secondly, a new multi-scale feature fusion module is designed to enhance the perception of feature information to supplement the positional information and high-resolution of the deep feature map. Finally, the SA attention mechanism is integrated into the neck network before the output of the feature map to enhance the correlation between the feature map channels and improve the fine-grained feature fusion ability of YOLOv5s. Experimental results show that compared with various YOLO algorithms, the proposed algorithm improves the detection accuracy and speed to a certain extent. Compared with the YOLOv3, YOLOv3-spp, YOLOv5s and YOLOv5m models, the average accuracy increases by 18.1%, 15.2%, 6.9% and 1.9%, respectively. It can effectively reduce the missed detection of occluded sperm and achieve lightweight and efficient multi-sperm target detection.

As a newly proposed video coding standard, Versatile Video Coding (VVC) has adopted some revolutionary techniques compared to High Efficiency Video Coding (HEVC). The multiple-mode affine motion compensation (MM-AMC) adopted by VVC saves approximately 15%-25% Bjøntegaard Delta Bitrate (BD-BR), with an inevitable increase of encoding time. This paper gives an overview of both the 4-parameter affine motion model and the 6-parameter affine motion model, analyzes their performances, and proposes improved algorithms according to the symmetry of iterative gradient descent for fast affine motion estimation. Finally, the proposed algorithms and symmetric MM-AMC flame of VTM-7.0 are compared. The results show that the proposed algorithms save 6.65% total encoding time on average, which saves approximately 30% encoding time of affine motion compensation.

The integrated circuit (IC) is a subject for which researchers need practical experience, but its experiment cost is high, the risk involved is high, and it is not easy to carry out experiments on a large scale. This paper designs a three-dimensional integrated circuit virtual experiment platform based on Unity3d. The platform uses the Unity3d and 3ds Max tools to build a three-dimensional model of instruments, equipment, electronic components, and ultra cleanroom laboratory scenes in an integrated circuit experiment. In addition, it uses C# script to develop the functions and three-dimensional simulation of general virtual instruments and equipment and deploys the experimental website using the frame of the jspxcms open source website. The experimental platform arranges the three-dimensional web file WebGL file in the cloud. Students can use video and text materials to acquire basic IC knowledge at any time and conduct IC virtual experiments safely, efficiently, and without constraints. At present, the platform has been tested and used in teaching and has received high praise and recognition from students.

Sperm concentration and motility are generally analysed only in the discrete state in microscopic videos. As for sperm nonspecific aggregation areas, it brings difficulties to accurate sperm detection. In this paper, an algorithm for nonspecific aggregates automatic segmentation, detection and tracking of sperm is proposed. A grid model commensurate with the size of a sperm head is created to segment nonspecific aggregation areas. Multi‐scale edge function and new energy functional are designed based on the level set method to realize sperm head segmentation. In the sperm tracking stage, we improve the weight condition and the standard of trust flow quantization based on graph theory method, and simplify the sperm tracking to the vertex matching between two frames to solve the matching failure problem of adjacent frames with small space distance. The proposed method achieves accurate segmentation of sperm non‐specific aggregation regions, which outperforms the level set methods of LBF and SBGFR. At the same time, our method can real‐time calculation of sperm concentration and motility during sperm tracking. It is compared with four state‐of‐the‐art algorithms, and it gives lower tracking error rates, which has potential applications in male fertility field.

With the rapid development of mobile Internet technology and artificial intelligence technology, the digital publishing industry is in urgent need of using intelligent technology to change the current way of content production and service. Most of the e-book resources owned by publishing enterprises are in PDF format, which is not suitable for reading on mobile devices, and it is not convenient to directly extract key information and construct knowledge graph. With this in mind, this article designs a PDF automatic indexing scheme that can identify all the element information in PDF and output structured data automatically and then extract all the key information in it to generate a keyword library with tag weights. The scheme mainly involves two key technical points: parsing PDF based on text features and grammar rules and extracting keywords based on tag weights. The former visualizes the text block in PDF into a rectangular area, divides the elements by clustering algorithm, and, finally, outputs structured data containing all the information. The latter combines the tags and their weights in the structured data and extracts the keywords in it by the inter-word relation algorithm. The structured data and keywords database produced by this scheme can be used to produce intelligent e-book and build knowledge graph, thus helping publishing enterprises to transform from a content service provider to an intelligent knowledge service provider. This transformation can deeply excavate the core value of the content held by the publishing industry and promote the digitization and intelligentization process of the whole industry.

A low-Earth-orbit (LEO) satellite network can provide full-coverage access services worldwide and is an essential candidate for future 6G networking. However, the large variability of the geographic distribution of the Earth's population leads to an uneven service volume distribution of access service. Moreover, the limitations on the resources of satellites are far from being able to serve the traffic in hotspot areas. To enhance the forwarding capability of satellite networks, we first assess how hotspot areas under different load cases and spatial scales significantly affect the network throughput of an LEO satellite network overall. Then, we propose a multi-region cooperative traffic scheduling algorithm. The algorithm migrates low-grade traffic from hotspot areas to coldspot areas for forwarding, significantly increasing the overall throughput of the satellite network while sacrificing some latency of end-to-end forwarding. This algorithm can utilize all the global satellite resources and improve the utilization of network resources. We model the cooperative multi-region scheduling of large-scale LEO satellites. Based on the model, we build a system testbed using OMNET++ to compare the proposed method with existing techniques. The simulations show that our proposed method can reduce the packet loss probability by 30% and improve the resource utilization ratio by 3.69%.