As a fundamental assumption in simultaneous localization and mapping, the static scenes hypothesis can be hardly fulfilled in applications of indoor/outdoor navigation or localization. Recent works about simultaneous localization and mapping in dynamic scenes commonly use heavy pixel-level segmentation net to distinguish dynamic objects, which brings enormous calculations and limits the real-time performance of the system. That restricts the application of simultaneous localization and mapping on the mobile terminal. In this article, we present a lightweight system for monocular simultaneous localization and mapping in dynamic scenes, which can run in real time on central processing unit (CPU) and generate a semantic probability map. The pixel-wise semantic segmentation net is replaced with a lightweight object detection net combined with three-dimensional segmentation based on motion clustering. And a framework integrated with an improved weighted-random sample consensus solver is proposed to jointly solve the camera pose and perform three-dimensional object segmentation, which enables high accuracy and efficiency. Besides, the prior information of the generated map and the object detection results is introduced for better estimation. The experiments on the public data set, and in the real-world demonstrate that our method obtains an outstanding improvement in both accuracy and speed compared to state-of-the-art methods.

This paper describes a study that built a neural network prediction model based on feature extraction, focusing on text analysis and image analysis of WeChat official accounts reading quantity. Based on the embedding method of the deep learning model, we extracted the text features in the title and the image features in the cover picture, explored the relationship between these features and the reading quantity, and built a neural network model based on these features to predict the reading quantity. The results show that there is a phenomenon of sentiment fusion in the text, and a sentence vector model based on Doc2Vec and a neural network model both had a good performance. This paper proposes a tool that can predict the reading quantity in advance and help administrators adjust the titles and images according to the predicted results. INDEX TERMSFeature extraction; neural network; WeChat official accounts; Doc2Vec; user engagement I. INTRODUCTIONSocial media platforms, such as Twitter and Facebook, provide opportunities for people to create, communicate, and share ideas. In China, WeChat is a social media platform with strong communication and influencing characteristics. Administrators apply for WeChat official accounts to publish different kinds of articles or news on the platform, and readers can obtain and share information. By November 2017, WeChat had gathered more than 10 million official accounts, including 3.5 million monthly active official accounts and 797 million monthly active users [1]. Many authors have formed their own brands through original articles and become entrepreneurs on WeChat.Previous research on WeChat has focused on user behaviors and attitudes as well as the influence mechanism and communication power of WeChat as a social media platform. Specifically, it involves user satisfaction, user attitude, user intention [2-5], user engagement behavior [6][7], and the influence mechanism and effect of WeChat as an information communication platform on service provided by users [8][9][10][11]. However, there has been little research on deeper mining and exploration of the text through natural language processing (NLP). Moreover, as far as we know, there has been little research on the analysis of the cover image of WeChat official accounts. Our study focuses on text analysis and image analysis of WeChat official accounts reading quantity, which contributes to the research in this specific field and addresses

In today's state‐of‐the‐art high‐efficiency silicon solar cells need to be inserted a thin insulating layer in order to reduce the recombination losses between the carrier transport layer and Si surface, which can form a tunneling junction (TJ), thus increasing the performance of the TJ solar cells comparable with the pn junction structure. However, the copper indium gallium selenium (CIGS) solar cells inevitably lead to the losses of the carrier transport due to the interface which is widely assumed as the pn‐heterojunction. Herein, the TJ solar cells, aiming to enhance the performance of the solar cells, are fabricated by inserting the TiO2 between CIGS/CdS interface deposited by atomic layer deposition (ALD). By inserting the TiO2 insulating layer, the CIGS/TiO2/CdS structure can be effectively reduced the interface recombination, which leads to a reduced band bending in the p‐CIGS surface and compromises its field‐effect passivation. As a result, the CIGS solar cell with the tunneling junction achieves the 15.57% based on the stainless steel (SS) substrate, by introducing a barrier layer and doping NaF. These results provide an important preliminary foundation for the development of the CIGS solar cells with the tunneling junction structure.