Search

Due to the variable shape and size of the liver in abdominal CT and the complex surrounding tissues, accurate liver segmentation in CT remains a challenge. In the field of two-dimensional image analysis, the squeeze-excitation (SE) method effectively adjusts the input feature information. Our work is based on SE method to obtain feature information that is more relevant to the liver region through improved feature recalibration methods. This work adds the MSCR block, which can do space and channel feature recalibration, to 3D full convolution network. The 3D space part of the MSCR block re-calibrates the feature of liver voxel space. And in the channel recalibration part, it changes the original spatial compression operation according to the none-local method, so the spatial sampling results contain more global context information. Multiple metrics are used to evaluate the proposed model on the LiTS dataset, and it achieves better segmentation performance than other comparison models such as Attention U-Net. The method is further tested on the 3DIRCADb dataset, which proves its effectiveness and stability. Thus, the model proposed in the paper is effective for improving the performance of liver segmentation.

In every organization the data is a significant part that can be separated as structured, unstructured and semi-structured. Unstructured data cannot be administered in the real-time by RDBMS or Hadoop. PySpark can be used for realtime data analysis of movie rating data collection. To generate the modified recommendations, method is intended that is Recommender Systems. Some complications that a Recommender System encountered are data sparsity, cold start, popular bias and scalability. In this research paper collaborative technique is applied for m-ovie rating recommender system. By getting benefits of collaborative filtering technique, we conclude that we use perspective metadata that is freely accessible. Concealed suggestions occur in the middle of movie ratings to personalize recommendations. Our recommended methodology demonstrate novelty by providing the modified recommendations irrespective of research field and of customer’s capability. By the usage of MovieLens dataset a significant development above supplementary standard techniques in evaluating global enactment and capability to yield appropriate and top-ratings at the top of the recommendation list.

Aiming at the centralization problem of the traditional medical device traceability system. This paper proposes a medical device traceability system based on blockchain technology. By integrating the traditional tracing system with the blockchain technology, based on the decentralization and non-tamperability of the blockchain, the alliance chain and the smart contract construction system are adopted. Manufacturers, distributors, hospitals and consumers in the system can access product information and trace the entire circulation process of the product through RFID. In addition, consumers do not need to register an account number, but directly through the RFID query products related to the main information, such as the shelf life, production enterprises, production time, product price, etc. This system solves the problem that the traditional system is vulnerable to network attack due to centralization, which can improve the robustness of the system, enhance the reliability of data and realizes the traceability of the production, sales and use of medical devices. Therefore, it ensures that the source of medical devices can be traced, the flow can be tracked, the information can be inquired and the responsibility can be investigated to ensure the public consumption safety.

With the rapid development of Internet technology, various social networking platforms, especially mobile social networking platforms, continue to increase, resulting in a large amount of public opinion information. Internet public opinion has a clear emotional orientation, and its emotional orientation is very easy to spread and be infected, and even affect the development of the event. Aiming at the characteristics of lyric information rich and which are easy to change with time, the lyric theme analysis model and the lyric emotion evolution model are proposed. The LDA model is used to extract the topic from the lyric text in a period of time, and the sensational heat value is calculated according to the forwarding amount and the number of comments, and the lyrical theme with the highest heat is obtained. The relative entropy between sub-topics in the adjacent time slice of a specific hot topic is calculated, and the degree of association between the topics in the adjacent time slice is determined, thereby analyzing whether there is a split of the sub-topic and a new topic. Then the evaluation object is extracted, combined with the joint deep neural network model to judge the emotions of each evaluation object in different time, and the emotional evolution of the hot topic is analyzed from multiple dimensions. Finally, an example analysis of the network public opinion information from June to July 2018 is carried out to verify the validity of the above model. The model effectively solves the problems of immature emotion analysis model and low accuracy of emotion classification in the current public opinion analysis.

In recent years, deep convolutional neural networks have gradually become the preferred method for image processing. After the development of Classification, Detection and Segmentation, a large variety of state-of-the-art models and algorithms have emerged in the field. However, for some specific data sets or tasks, not all methods are applicable, which is inconvenient to researchers. This paper took the data set provided in the airbus ship detection challenge in Kaggle as an example to explore an easy and effective method for segmentation tasks of data sets with class imbalance. This paper used U-Net with a pre-trained ResNets model, and tried different methods to explore the feature of the set. In the process of training ResNets, this paper proposed a new convolutional block structure which is inspired by Fibonacci sequence, but the effect is not good. In the end, the mF2 values of the models this paper trained achieved good results, which is better than the model of the combined training of ResNets and ordinary U-Net34. Moreover, the training parameters are less than that. This paper believe that this simple and effective training method will bring convenience to researchers in related fields.