Semantic Expression Research Articles

FACTITIVE CAUSATIVE CONSTRUCTIONS IN MODERN FRENCH (BASED ON THE MATERIAL OF THE NOVEL BY L. SLIMANI «THE COUNTRY OF OTHERS»)

Statement of the problem. Causation is one of the leading ways of expressing subject-object relations, as well as causal relationships in the linguistic and extralinguistic environment. This article discusses the lexical and grammatical category of the causative and the means of its representation in modern French. Particular attention is paid to the study of causative constructions as the leading and most frequent way of updating the causative meaning in French. Results. As part of the study, the semantic classification of the verbs with a causative meaning was studied, including factitive, permissive and assistive types. Due to the thematic conditionality of this work, the semantic features of the causative constructions of the factitive type, presented in the text of the novel by the modern French author L. Slimani «The country of others», in particular its first («War, war, war») and the second part («Watch us dance»). Based on the analysis carried out during the study, 211 constructions with factitive causation were identified in both parts of the novel. It was possible to establish that the construction faire + inf. (to force to do something) is most often used: «War, war, war» – 36%, «Watch us dance» –43%. Also, the construction ordonner de + inf. (to order to do something), the percentage of which was: «War, war, war» – 25%, «Watch us dance» –18%, was also second in frequency of use. Conclusion. The results obtained confirm the opinion of some gallist scientists that faire + inf. is the most frequent example of the expression of factitive causative semantics in French. However, in addition to the named construction with the verb faire, in both parts of the novel the construction with the verb ordonner is repeatedly used.

The Evaluation of Pseudo-Kufic in Military Weapons and Clothing Depictions

The subject of the study was the military weapons in the depictions and the pseudo-kufic in clothes. The pseudo-kufic, which were seen in Byzantine and Western art in the 9th century, and which were applied in Christian religious structures as of the 10th century, were evaluated by scientists as a purely symbolic form of semantic and visual expression that was borrowed from a foreign language that could not be read. The dual nature of pseudo-kufic, in which Arabic names, verses, good intentions, and wishes were conveyed, allowed them to spread in a wide range and be evaluated as a transitional decoration at the crossroads of various fields. It was reported previously that the practices explained with various expressions such as trade, conquest, occupation, captives, and interaction had an essentially apotropaic meaning, and those that could not be read were ornaments that had decorative designs. It was determined that pseudo-kufic in depictions of military weapons and clothing had a special meaning, and therefore, they were frequently applied in Byzantine painting, illustrated manuscripts, and small handicrafts. The Arabic independent letters in the depictions of military clothes and weapons, the texts in the form of the abbreviation of the verse Al-Mulk or Al-Mulk Lillah had a U-form or mirror form structure. The application of the verse Al-Mulk or Al-Mulk Lillah on military weapons and clothes depictions in practice shows that the expression of power was at the forefront rather than the meaning of property in buildings and other works.

The current state of radiodiagnostics of degenerative changes in the lumbar spine (literature review)

Examination of the spine using radiological methods remains the most frequently conducted study in the outpatient practice of almost any radiologist. In most cases, changes in the spine are degenerative and dystrophic in nature. These changes in the spine are the leading cause of loss of activity among adults and the elderly, and cover a wide range of age-related structural changes. It is of great importance to understand the possibilities and limitations of radiological methods of diagnostics. The description of morphological changes observed in degenerative and dystrophic changes requires the use of unified terminology and classifications among clinicians and radiologists.The aim. To present modern concepts in the assessment of degenerative changes of the spine using radiological methods. A description of the standardized international nomenclature of intervertebral disc pathology and current classifications of spinal canal stenosis are presented.Material and methods. For literature search, we used electronic databases MEDLINE (PubMed), eLibrary, EMBASE and Cochrane Library with a selection of sources published from 2000 to 2021. We analyzed the works devoted to the diagnosis of degenerative changes in lumbar spine and to degenerative stenosis.Conclusion. The key to productive communication between physicians is the uniformity or standardization of terminology and definitions used. It is important that the protocol of description, its terms, semantic expressions be uniform and understandable to specialists who are involved in the diagnosis and treatment of the spine. Magnetic resonance imaging is a valuable method in the diagnosis of degenerative changes of the spinal column, but it has its limitations.

Effects of syntactic structure on the processing of lexical repetition during sentence reading.

Previous research has demonstrated that the ease or difficulty of processing complex semantic expressions depends on sentence structure: Processing difficulty emerges when the constituents that create the complex meaning appear in the same clause, whereas difficulty is reduced when the constituents appear in separate clauses. The goal of the current eye-tracking-while-reading experiments was to determine how changes to sentence structure affect the processing of lexical repetition, as this manipulation enabled us to isolate processes involved in word recognition (repetition priming) from those involved in sentence interpretation (felicity of the repetition). When repetition of the target word was felicitous (Experiment 1), we observed robust effects of repetition priming with some evidence that these effects were weaker when repetition occurred within a clause versus across a clause boundary. In contrast, when repetition of the target word was infelicitous (Experiment 2), readers experienced an immediate repetition cost when repetition occurred within a clause, but this cost was eliminated entirely when repetition occurred across clause boundaries. The results have implications for word recognition during reading, processes of semantic integration, and the role of sentence structure in guiding these linguistic representations.

ПРОБЛЕМИ ВРАХУВАННЯ КООПЕРОВАНОЇ ФОРМИ ЕКСПЛУАТАЦІЇ ПРИ ФОРМУВАННІ СПОРТИВНИХ КОМПЛЕКСІВ ЗАКЛАДІВ ВИЩОЇ ОСВІТИ

This article is devoted to the issue of cooperative operation of sports facilities of educational institutions by residents of nearby residential areas. Our interest in this question is based on the problem of low involvement of the population of Ukraine in physical culture and sports, as well as the lack of capacity of the existing network of sports facilities of the country. Considered several scientific papers in which the question of cooperative operation of sports facilities of educational institutions by residents of nearby areas. It is noted that the sports facilities of educational institutions make up a large part of the network of sports facilities of the country. However, most of the existing educational institutions sporting facilities have been built without taking into account the possibility of use by different groups of people. The article considers several characteristic features of such sports facilities that make their use by residents of nearby areas inconvenient. Both the functional aspect (the location of the object in the back of the site; location of sports halls on the upper or basement floors of educational buildings, etc.), and the semantic (expressiveness and informativeness of the environment) and aesthetic aspects are considered. The question of the motivational capacity of higher education institutions' sports complexes is raised. The modern state of a network of sports facilities at higher educational institutions is considered. The question of the negative influence of their obsolete and in need of repair condition on the motivation of visitors and the effectiveness of the process of physical education on their basis is highlighted. Insufficient attention to the issue of cooperative operation of the sports complex of higher educational institutions by other groups of the population is noted. In particular, the question of taking this form of operation into account when calculating the simultaneous capacity of a higher education institution's sports complex remains unclear. This issue requires further study.

Classification of Learning Sentiments of College Students Based on Topic Discussion Texts of Online Learning Platforms

Depicting the online learning process of student users from multiple angles can help implement deep learning and effectively improve their online learning quality, and it’s a practical and very meaningful work to mine the data burying in the topic discussion texts of online learning platforms so that useful information could be extracted and attained to help teachers better understand students’ learning sentiments and assist students to know of the learning status of their peers. However, in existing conventional sentiment analysis methods, the sample data with uncategorized tags are still labelled manually, and such work is usually time consuming and inefficient. In view of these defects, this paper aims to study the classification of college students’ learning sentiments based on the topic discussion texts of online learning platforms. In the beginning, this paper gave the overall structure of the proposed college student Learning Sentiment Classification (LSC) algorithm, and discussed the similarity between the topic discussion content and the teaching content. Then, this paper proposed to integrate Convolution Neural Network (CNN) with the Long-Short Term Memory (LSTM) network to build the said LSC model, so as to merge the advantages of the two and improve the accuracy of learning sentiment rating. After that, embedding layers of static words and non-static words were introduced into the proposed model for the purpose of realizing the mining of specific textual information while enhancing the semantic expression ability of the words. At last, experimental results verified the effectiveness of the proposed model.

LitCovid ensemble learning for COVID-19 multi-label classification.

The Coronavirus Disease 2019 (COVID-19) pandemic has shifted the focus of research worldwide, and more than 10 000 new articles per month have concentrated on COVID-19-related topics. Considering this rapidly growing literature, the efficient and precise extraction of the main topics of COVID-19-relevant articles is of great importance. The manual curation of this information for biomedical literature is labor-intensive and time-consuming, and as such the procedure is insufficient and difficult to maintain. In response to these complications, the BioCreative VII community has proposed a challenging task, LitCovid Track, calling for a global effort to automatically extract semantic topics for COVID-19 literature. This article describes our work on the BioCreative VII LitCovid Track. We proposed the LitCovid Ensemble Learning (LCEL) method for the tasks and integrated multiple biomedical pretrained models to address the COVID-19 multi-label classification problem. Specifically, seven different transformer-based pretrained models were ensembled for the initialization and fine-tuning processes independently. To enhance the representation abilities of the deep neural models, diverse additional biomedical knowledge was utilized to facilitate the fruitfulness of the semantic expressions. Simple yet effective data augmentation was also leveraged to address the learning deficiency during the training phase. In addition, given the imbalanced label distribution of the challenging task, a novel asymmetric loss function was applied to the LCEL model, which explicitly adjusted the negative-positive importance by assigning different exponential decay factors and helped the model focus on the positive samples. After the training phase, an ensemble bagging strategy was adopted to merge the outputs from each model for final predictions. The experimental results show the effectiveness of our proposed approach, as LCEL obtains the state-of-the-art performance on the LitCovid dataset. Database URL: https://github.com/JHnlp/LCEL.

Evaluation Cloud Model of Spontaneous Combustion Fire Risk in Coal Mines by Fusing Interval Gray Number and DEMATEL

Coal still occupies a key position in China’s energy consumption structure, and ensuring safe production in coal mines is a key focus for ensuring energy security. Spontaneous combustion fires in coal mines are a serious threat to the sustainability of safe production in coal mines. In order to prevent coal mine fire risk scientifically and effectively and to assess the level of disaster risk effectively and rationally, a study was conducted on the risk of spontaneous combustion fires in underground coal mines. An evaluation cloud model of spontaneous combustion fire risk in coal mines integrating the interval gray number with the Decision-Making Trial and Evaluation Laboratory (DEMATEL) was established. Seventeen representative risk evaluation indicators were selected, and a coal mine spontaneous combustion fire risk evaluation index system was constructed based on four aspects: personnel, machinery, environment, and management. The interval gray number theory was introduced to improve the classical DEMATEL analysis method, which fully expresses the expert empirical knowledge and solves the problem of ambiguity and randomness in the semantic expression of expert evaluation. The relative importance of each indicator was determined by analyzing the influence relationships between risk evaluation indicators through the improved DEMATEL. A cloud model capable of transforming quantitative descriptions and qualitative concepts was used for comprehensive evaluation of risk, and based on the results of DEMATEL analysis, a comprehensive evaluation cloud model of coal mine spontaneous combustion fire risk was formed. Finally, the validity and practicality of the model were verified by using a mine in Shenmu City, Shaanxi Province, China as an example. This study provides a powerful tool to prevent spontaneous combustion fires in coal mines and makes a positive contribution to the sustainable development of coal mine safety management.

A novel feature integration and entity boundary detection for named entity recognition in cybersecurity

Owing to continuous cyberattacks, a large amount of threat intelligence is generated online every day. However, threat intelligence is mostly unstructured and multisource heterogeneous text. It is difficult for security analysts to understand the implicit threat in time. Knowledge Graph (KG) is an important research topic in recent years, which can perform automated and real-time analysis of threat intelligence in cybersecurity. As one of the critical technologies of KG, named entity recognition (NER) can identify cyberattack-related entities. It has been proved that long-distance structured information captured by dependency trees provides a rich semantic expression for the neural network. However, existing research works are more focused on the simple linear stack of neural networks when utilizing structured features. The interaction between different types of neural networks is vague. In addition, the existing models are insensitive to the boundaries of complex entity terms in cybersecurity. In this study, we propose a new feature integration and entity boundary detection (FIEBD) model. In our model, a new pretrained language model, PERT, is applied to obtain word embedding of cyber texts. Moreover, a novel neural network cell, namely GARU, is developed to incorporate different types of features extracted from graph neural networks and recurrent neural networks. It combines the graph encoder with the gate mechanism, aiming to obtain better hidden representation by explicit interaction. Furthermore, considering a large number of complex entities in cybersecurity, we contribute an entity boundary detection module to perform entity head and tail prediction as an augmentation task. We conduct extensive experiments on cybersecurity datasets. The results demonstrate that the proposed model achieves better performance than existing baseline methods.

Deep feature extraction of retinal images for diabetic retinopathy enhanced diagnosis

Most people of working age groups globally around are affected by diabetic retinopathy, which is the most common cause of preventable vision impairment. Recent studies have improved our comprehension of the need for more effective and affordable methods of identifying, managing, diagnosing, and treating retinal disease in clinical eye care practice. The development of computer-aided diagnosis tools must consider the significance of diabetic retinopathy screening programmes and the challenge of obtaining a valid early diagnosis of diabetic retinopathy at an affordable price. Retinal image analysis using computer-aided disease diagnosis may make it easier to screen the population for diabetes mellitus on a mass scale and may also assist physicians to make better use of their time. The creation of image processing models using big data analysis may run into many technical challenges, including the visualization analysis needed by image processing technologies, the semantic expression, and storage of large image samples, the complexity of the algorithms needed for feature extraction, recognition, and prediction analysis using big data, as well as the time and memory requirements.

MaCOnto: A robust maize crop ontology based on soils, fertilizers and irrigation knowledge

• A robust OWL rule based autonomous evolution ontology for representing the maize crop knowledge. • MaCOnto: implemented as rules via SWRL, and queried by SQWRL. • Novel algorithmic design for the ontology to autonomously evolved is implemented in java environment. • MaCOnto capable to infer and respond to user's query based on its contextual information. The demand for relevant information in a timely manner portrays the significance of knowledge management in all areas of lives; for instance, agriculture. To this end, soils, fertilizers and irrigation as agronomic concepts are essential knowledge inputs for any crops, such as maize. Conversely, there is always difficulty in timely retrieval of these relevant information owing to the unstructured nature of data in repositories, and complexity of concepts mismatch. Sequel to this development, ontology, a semantic data modeling technique is promising as it has been recently employed to deal with these challenges across different domains. However, the robustness of ontology, in terms of semantic expressivity of hidden knowledge, and autonomous growth of ontology leave some gaps to contend with. In view of this development, this research aims to design a robust OWL Rule based ontology for maize crop domain by considering primarily soils, fertilizers and irrigation agronomic concepts capable to evolve autonomously. The proposed ontology herein christened MaCOnto , is developed using the adapted six steps ontology-engineering principle. Over 1,430 entities are encoded in OWL; eighty Competency Questions (CQs) validated by domain experts are modeled in FOL, and implemented as rules via SWRL. Thus, the ontology is queried by SQWRL. Besides, the novel algorithmic design for the ontology to autonomously evolve is implemented in Java environment by employing WordNet. The results obtained from structural based evaluation show an outstanding performance across the eight metrics. Similarly, the results of the competency-based evaluation are also promising. Therefore, the proposed MaCOnto is a robust application based ontology capable to infer and responds to user's query based on its contextual information.

Semantic Modeling of Ship Behavior in Cognitive Space

Ship behavior is the semantic expression of corresponding trajectory in spatial-temporal space. The intelligent identification of ship behavior is critical for safety supervision in the waterborne transport. In particular, the complicated behavior reflects the long-term intentions of a ship, but it is challenging to recognize it automatically for computers without a proper understanding. For this purpose, this study provides a method to model the behavior for computers from the perspective of knowledge modeling that is explainable. Based on our previous work, a semantic model for ship behavior representation is given considering the multi-scale features of ship behavior in cognitive space. Firstly, the multi-scale features of ship behavior are analyzed in spatial-temporal dimension and semantic dimension individually. Then, a method for multi-scale behaviors modeling from the perspective of semantics is determined, which divides the behavior scale into four sub-scales in cognitive space, considering spatial and temporal dimensions: action, activity, process, and event. Furthermore, an ontology model is introduced to construct the multi-scale semantic model for ship behavior, where behaviors with different semantic scales are expressed using the functions of ontology from a microscopic perspective to a macroscopic perspective consecutively. To validate the model, a case study is conducted in which ship behavior with different scales occurred in port water areas. Typical behaviors, which include leveraging the axioms expression and semantic web rule language (SWRL) of the ontology, are then deduced using a reasoner, such as Pellet. The results show that the model is reasonable and feasible to represent multi-scale ship behavior in various scenarios and provides the potential to construct a smart supervision network for maritime authorities.

MANIFESTATIONS OF MYTHS IN ARCHITECTURAL ENVIRONMENT

The article examines the aspects of myth, perceived as an existential, reality-based narrative and a means of semantic expression, which are associated with architectural theory and the expression of the architectural environment. The question about the role myths play in the process of adding meaning to a public space is raised. The article also addresses the importance of place recognition and the social and psychological need it creates to encode symbols in the architectural environment. Specific cases of myth coding are discussed and features of their expression are distinguished.

Classification of Land Cover Remote-Sensing Images Based on Pattern Recognition

With the development of remote sensing technology, remote sensing image data plays an active role in the dynamic monitoring of global resource changes and land cover utilization. Remote sensing image land cover classification is an important application direction of remote sensing data; how to further improve the accuracy of remote sensing image land cover classification is very important for the effective application of remote sensing data. The traditional remote sensing image land cover classification is mainly to classify remote sensing data according to the spectral data of ground objects. However, due to the complex environment of remote sensing images and the dynamic changes of the environment, traditional classification methods based only on pixel spectral data are often unable to achieve. A satisfactory classification result is achieved. In addition, some researchers have also proposed to combine pixel neighborhood texture information to supplement spectral feature data. Although the traditional classification method based on spectral features solves the problem of time-consuming visual interpretation, to a certain extent, due to the limited semantic expression ability and poor generalization ability of the design features, the classification accuracy is still not very satisfactory. This paper mainly studies the classification method of land cover remote sensing image based on pattern recognition. This paper is based on the experimental results of remote sensing data in Nanjing Yuhuatai District in 2018 and 2019. The ground resolution of the data is 2.5 meters. Data is projected, corrected, and equalized. Half of the images covering 43.75 square kilometers are used as training samples, and the remaining 50 square kilometers are used for detection. In the classification results of this IndianPines data, OA only increased by nearly 10% to 86.2%, AA increased by 13%, r was 82.77%, and Kappa coefficient was 0.84. In the classification results of Salinas data, both OA and AA increase by about 5%, and the optimization effect is not obvious.

Progressive Deep Non-Negative Matrix Factorization Architecture with Graph Convolution-based Basis Image Reorganization

Deep non-negative matrix factorization is committed to using multi-layer structure to extract underlying parts-based representation. However, the basis images obtained by continuous depth factorization is too sparse, resulting in too fragmented parts reflected by the basis image. This makes the number of factorization layers limited and the underlying local feature representation is inaccurate. Therefore, we propose a novel progressive deep non-negative matrix factorization (PDNMF) architecture that adds a basis image reconstruction step to the successive basis image factorization steps. This helps the basis image in depth factorization to maintain better robustness of feature representation. In the reconstruction step, the attribute similarity graph (ASG) is constructed to describe the semantic expression ability of each basis image. With the help of the ASG, the basis image enhances its own semantic integrity through graph convolution without drastically destroying its representation. The evaluation in image recognition shows that the recognition accuracy of the proposed PDNMF improves with the increase of layers. Our method outperforms the state-of-the-art deep factorization methods in image recognition.

DeepLocBIM: Learning Indoor Area Localization Guided by Digital Building Models

Fingerprinting-based indoor localization is a cost-effective approach to provide coarse-grained indoor location information for pedestrian mass-market applications without the requirement of installing additional positioning infrastructure. While most solutions aim at pinpointing the exact location of a user, estimating a zone/area is a promising approach to achieve a more reliable prediction. Area localization predominantly utilizes a predetermined building model segmentation to obtain zone/area labels for collected fingerprints. We propose a novel approach to multifloor indoor area localization by directly predicting polygon zones that contain the position of the user. Our model learns to construct the zones from the wall segments and thus predicted areas have a high conformity to the underlying building model (semantic expressiveness). On a self-collected as well as on a public fingerprinting data set, we compare our model with two reference approaches. We demonstrate that the utilized surface areas of the polygons are on average up to 50% smaller than those of the reference models and provide a high semantic expressiveness without requiring manual floor plan segmentation.

Sentiment Analysis of Chinese Reviews Based on BiTCN-Attention Model

It is of great significance for individuals, enterprises, and government departments to analyze and excavate the sentiment in the comments. Many deep learning models are used for text sentiment analysis, and the BiTCN model has good efficacy on sentiment analysis. However, in the actual semantic expression, the contribution of each word to the sentiment tendency is different, BiTCN treats it fairly and does not pay more attention to the key sentiment words. For this problem, a sentiment analysis model based on the BiTCN-Attention is proposed in this paper. The Self-Attention mechanism and Multi-Head Self-Attention mechanism are added to BiTCN respectively to form BiTCN-SA and BiTCN-MHSA, which improve the weight of sentiment words and the accuracy of feature extraction, to increase the effect of sentiment analysis. The experimental results show that the model accuracies of BiTCN-SA and BiTCN-MHSA in the JingDong commodity review data set are 3.96% and 2.41% higher than that of BiTCN, respectively. In the comment data set of DianPing, the accuracy of BiTCN-SA and BiTCN-MHSA improved by 4.62% and 3.49%, respectively, compared with that of BiTCN.

Dependency graph enhanced interactive attention network for aspect sentiment triplet extraction

Aspect sentiment triplet extraction is an extremely daunting task designed to identify the triplets from comments, where each triplet is composed of an aspect term, the related opinion term, and the sentiment between them. Existing research efforts majorly construct a novel tagging scheme to avoid the disadvantages of pipeline methods. However, the improvement is limited due to neglecting the implicit grammatical relationships among the three elements in a triplet. To cope with this limitation, we put forward an innovative Dependency Graph Enhanced Interactive Attention Network, which explicitly introduces the syntactic and semantic relationships between words. Specifically, an interactive attention mechanism is conceived to jointly consider both the contextual features learned from Bi-directional Long Short-Term Memory and the syntactic dependencies learned from the correspondent dependency graph in an iterative interaction manner. In addition, we notice that words with different Part-of-Speech categories have different contributions to the semantic expression of sentences. Accordingly, the information of different Part-of-Speech categories is recognized during the modeling process to properly capture the semantic relationships. Experiments on the benchmark datasets originally derived from SemEval Challenges illustrate that our presented approach has superiority over strong baselines.

Sociocultural identity: judgments, definitions and modern problems

In historical and philosophical knowledge, since antiquity, the content side of the concept of "identity" has been considered in line with the general philosophical category of "identity". In this article, the object of research is the issues of identity in relation to the socio-cultural sphere, while taking into account the understanding of identity from the standpoint of logic, ontology, sociology, anthropology, psychology, linguistics. Accordingly, the subject of this study is the phenomenon of socio-cultural identity, in particular, in the context of the construction of a new self-consciousness in a changing Russia. Identity is interpreted by the authors as the socio-psychological stability of the subjects of society from the individual to the social community at the level of an ethnic group and people. At the same time, special emphasis is placed on identifying the meanings of such components of identity as the presence of a world of spiritual values, the peculiarities of the development of the ethno-cultural factor of society and others in relation to the history and current state of Russian society. The political and legal context of the identity of modern Russian society is analyzed separately. Since language is one of the leading factors in the construction of (ethnic, social, religious, etc.) identity, the authors analyze sociolinguistic and linguoculturological processes affecting the formation of self-identification of the individual and society. In the concluding part of the article it is indicated that in the modern conditions of the development of the cultural and civilizational process, the formation of self-consciousness and socio-cultural identity become the dominant measure of the existence of modern society. Socio-cultural identity by its nature encompasses the spiritual, psychological, political, legal and moral atmosphere of society. According to the authors of the article, in relation to Russian society, the understanding of socio-cultural identity as a way of establishing one's own civilizational specificity and originality has acquired particular importance. The awareness of the Russian person of his own self as a fragment of modern civilization is acquiring a semantic expression of eventfulness today. The scientific novelty of this study consists in identifying the main characteristics of the socio-cultural identity of modern Russian society.

Semantic translation of face image with limited pixels for simulated prosthetic vision

Facial perception and cognition are among the most critical functions of retinal prostheses for blind people. However, owing to the limitations of the electrode array, simulated prosthetic vision can only provide limited pixel images, which seriously weakens the image semantics expression. To improve the intelligibility of face images with limited pixels, we constructed a face semantic information transformation model to transform real faces into pixel faces based on the analogy between human and artificial intelligence, named F2Pnet (face to pixel networks). This is the first attempt at face pixelation using deep neural networks for prosthetic vision. Furthermore, we established a pixel face database designed for prosthesis vision and proposed a new training strategy for generative adversarial networks for image-to-image translation tasks aiming to solve the problem of semantic loss under limited pixels. The results of psychophysical experiments and user studies show that the identifiability of pixel faces in characteristic and expression is much better than that of comparable methods, which is significant for improving the social ability of blind people. The real-time operation (17.7 fps) on the Raspberry Pi 4 B shows that F2Pnet has reasonable practicability.