Semantic Filtering Research Articles

A method for image–text matching based on semantic filtering and adaptive adjustment

As image–text matching (a critical task in the field of computer vision) links cross-modal data, it has captured extensive attention. Most of the existing methods intended for matching images and texts explore the local similarity levels between images and sentences to align images with texts. Even though this fine-grained approach has remarkable gains, how to further mine the deep semantics between data pairs and focus on the essential semantics in data remains to be quested. In this work, a new semantic filtering and adaptive approach (FAAR) was proposed to ease the above problem. To be specific, the filtered attention (FA) module selectively focuses on typical alignments with the interference of meaningless comparisons eliminated. Next, the adaptive regulator (AR) further adjusts the attention weights of key segments for filtered regions and words. The superiority of our proposed method was validated by a number of qualitative experiments and analyses on the Flickr30K and MSCOCO data sets.

Semantic Filtering and Source Coding in Distributed Wireless Monitoring Systems

Semantic Filtering and Source Coding in Distributed Wireless Monitoring Systems

Artificial intelligence auxiliary diagnosis and treatment system for breast cancer in developing countries.

In many developing countries, a significant number of breast cancer patients are unable to receive timely treatment due to a large population base, high patient numbers, and limited medical resources. This paper proposes a breast cancer assisted diagnosis system based on electronic medical records. The goal of this system is to address the limitations of existing systems, which primarily rely on structured electronic records and may miss crucial information stored in unstructured records. The proposed approach is a breast cancer assisted diagnosis system based on electronic medical records. The system utilizes breast cancer enhanced convolutional neural networks with semantic initialization filters (BC-INIT-CNN). It extracts highly relevant tumor markers from unstructured medical records to aid in breast cancer staging diagnosis and effectively utilizes the important information present in unstructured records. The model's performance is assessed using various evaluation metrics. Such as accuracy, ROC curves, and Precision-Recall curves. Comparative analysis demonstrates that the BC-INIT-CNN model outperforms several existing methods in terms of accuracy and computational efficiency. The proposed breast cancer assisted diagnosis system based on BC-INIT-CNN showcases the potential to address the challenges faced by developing countries in providing timely treatment to breast cancer patients. By leveraging unstructured medical records and extracting relevant tumor markers, the system enables accurate staging diagnosis and enhances the utilization of valuable information.

Train Distance Estimation for Virtual Coupling Based on Monocular Vision.

By precisely controlling the distance between two train sets, virtual coupling (VC) enables flexible coupling and decoupling in urban rail transit. However, relying on train-to-train communication for obtaining the train distance can pose a safety risk in case of communication malfunctions. In this paper, a distance-estimation framework based on monocular vision is proposed. First, key structure features of the target train are extracted by an object-detection neural network, whose strategies include an additional detection head in the feature pyramid, labeling of object neighbor areas, and semantic filtering, which are utilized to improve the detection performance for small objects. Then, an optimization process based on multiple key structure features is implemented to estimate the distance between the two train sets in VC. For the validation and evaluation of the proposed framework, experiments were implemented on Beijing Subway Line 11. The results show that for train sets with distances between 20 m and 100 m, the proposed framework can achieve a distance estimation with an absolute error that is lower than 1 m and a relative error that is lower than 1.5%, which can be a reliable backup for communication-based VC operations.

Hyperspectral image classification method based on semantic filtering and ensemble learning

The imbalance between limited training samples and extreme high spectral dimensions is a challenge for hyperspectral image classification. To significantly improve the classification of hyperspectral images in the case of small samples, we proposed a novel hyperspectral image classification method based on semantic filtering and ensemble learning. Semantic filtering synergistically combines the efficiency of the recursive filter and the effectiveness of the recent edge detector for scale-aware edge-preserving filtering, which can efficiently extract subjectively meaningful structures from natural scenes containing multiple-scale objects. Ensemble learning is used to fuse multi-scale features and to enhance the information-rich features. Experiments on three data sets prove that both semantic information and ensemble learning can improve classification accuracy, and their combination obtained the state-of-the-art performance.

Lite‐weight semantic segmentation with AG self‐attention

AbstractDue to the large computational and GPUs memory cost of semantic segmentation, some works focus on designing a lite weight model to achieve a good trade‐off between computational cost and accuracy. A common method is to combined CNN and vision transformer. However, these methods ignore the contextual information of multi receptive fields. And existing methods often fail to inject detailed information losses in the downsampling of multi‐scale feature. To fix these issues, we propose AG Self‐Attention, which is Enhanced Atrous Self‐Attention (EASA), and Gate Attention. AG Self‐Attention adds the contextual information of multi receptive fields into the global semantic feature. Specifically, the Enhanced Atrous Self‐Attention uses weight shared atrous convolution with different atrous rates to get the contextual information under the specific different receptive fields. Gate Attention introduces gating mechanism to inject detailed information into the global semantic feature and filter detailed information by producing “fusion” gate and “update” gate. In order to prove our insight. We conduct numerous experiments in common semantic segmentation datasets, consisting of ADE20 K, COCO‐stuff, PASCAL Context, Cityscapes, to show that our method achieves state‐of‐the‐art performance and achieve a good trade‐off between computational cost and accuracy.

Molecular cartooning with knowledge graphs.

Molecular "cartoons," such as pathway diagrams, provide a visual summary of biomedical research results and hypotheses. Their ubiquitous appearance within the literature indicates their universal application in mechanistic communication. A recent survey of pathway diagrams identified 64,643 pathway figures published between 1995 and 2019 with 1,112,551 mentions of 13,464 unique human genes participating in a wide variety of biological processes. Researchers generally create these diagrams using generic diagram editing software that does not itself embody any biomedical knowledge. Biomedical knowledge graphs (KGs) integrate and represent knowledge in a semantically consistent way, systematically capturing biomedical knowledge similar to that in molecular cartoons. KGs have the potential to provide context and precise details useful in drawing such figures. However, KGs cannot generally be translated directly into figures. They include substantial material irrelevant to the scientific point of a given figure and are often more detailed than is appropriate. How could KGs be used to facilitate the creation of molecular diagrams? Here we present a new approach towards cartoon image creation that utilizes the semantic structure of knowledge graphs to aid the production of molecular diagrams. We introduce a set of "semantic graphical actions" that select and transform the relational information between heterogeneous entities (e.g., genes, proteins, pathways, diseases) in a KG to produce diagram schematics that meet the scientific communication needs of the user. These semantic actions search, select, filter, transform, group, arrange, connect and extract relevant subgraphs from KGs based on meaning in biological terms, e.g., a protein upstream of a target in a pathway. To demonstrate the utility of this approach, we show how semantic graphical actions on KGs could have been used to produce three existing pathway diagrams in diverse biomedical domains: Down Syndrome, COVID-19, and neuroinflammation. Our focus is on recapitulating the semantic content of the figures, not the layout, glyphs, or other aesthetic aspects. Our results suggest that the use of KGs and semantic graphical actions to produce biomedical diagrams will reduce the effort required and improve the quality of this visual form of scientific communication.

Knowledge-Enhanced Dual-Channel GCN for Aspect-Based Sentiment Analysis

As a subtask of sentiment analysis, aspect-based sentiment analysis (ABSA) refers to identifying the sentiment polarity of the given aspect. The state-of-the-art ABSA models are developed by using the graph neural networks to deal with the semantics and the syntax of the sentence. These methods are challenged by two issues. For one thing, the semantic-based graph convolution networks fail to capture the relation between aspect and its opinion word. For another, minor attention is assigned to the aspect word within graph convolution, resulting in the introduction of contextual noise. In this work, we propose a knowledge-enhanced dual-channel graph convolutional network. On the task of ABSA, a semantic-based graph convolutional netwok (GCN) and a syntactic-based GCN are established. With respect to semantic learning, the sentence semantics are enhanced by using commonsense knowledge. The multi-head attention mechanism is taken to construct the semantic graph and filter the noise, which facilitates the information aggregation of the aspect and the opinion words. For syntactic information processing, the syntax dependency tree is pruned to remove the irrelevant words, based on which more attention weights are given to the aspect words. Experiments are carried out on four benchmark datasets to evaluate the working performance of the proposed model. Our model significantly outperforms the baseline models and verifies its effectiveness in ABSA tasks.

Microblog Sentiment Analysis Based on Dynamic Character-Level and Word-Level Features and Multi-Head Self-Attention Pooling

To address the shortcomings of existing deep learning models and the characteristics of microblog speech, we propose the DCCMM model to improve the effectiveness of microblog sentiment analysis. The model employs WOBERT Plus and ALBERT to dynamically encode character-level text and word-level text, respectively. Then, a convolution operation is used to extract local key features, while cross-channel feature fusion and multi-head self-attention pooling operations are used to extract global semantic information and filter out key data, before using the multi-granularity feature interaction fusion operation to effectively fuse character-level and word-level semantic information. Finally, the Softmax function is used to output the results. On the weibo_senti_100k dataset, the accuracy and F1 values of the DCCMM model improve by 0.84% and 1.01%, respectively, compared to the best-performing comparison model. On the SMP2020-EWECT dataset, the accuracy and F1 values of the DCCMM model improve by 1.22% and 1.80%, respectively, compared with the experimental results of the best-performing comparison model. The results showed that DCCMM outperforms existing advanced sentiment analysis models.

Software for Mapping and Extraction of Building Land Remote Sensing Data Based on BIM and Sensor Technology

In order to solve the problem of complex extraction caused by large feature dimension of remote sensing data, this paper proposes a dimension compression extraction method of urban building land remote sensing data under BIM Technology. Firstly, the remote sensing data is imported into the BIM model for lightweight processing to obtain the element information required for urban construction land and then analyze the urban construction land data, extract the key elements of BIM Technology through semantic filtering, and use the triangulation method to transform the remote sensing data into the triangulation model that can be processed by GIS model. Finally, the random projection method is used to reduce the dimension and compress the remote sensing data, and the remote sensing data extraction of urban construction land is realized through dictionary learning, vocabulary coding, and feature extraction. The experimental results show that the accuracy of extracting different land use types by this method is more than 99%, while the accuracy of extracting different land use types by depth learning method and PLS method is less than 98.5%. In addition, the signal-to-noise ratio of the image extracted by this method is significantly higher than that by depth learning method and PLS method. Conclusion. This method can effectively compress and extract the urban construction land in the remote sensing data, and the extraction accuracy of remote sensing data is high. It provides a technical basis for the approval of urban construction planning. It has the advantages of simple feature extraction and effective differentiation of ground objects.

Dimensionality Reduction and Extraction of Engineering Remote Sensing Data Based on Building Information Modeling and Geographical Information System

The high dimensionality of the modern remote sensing data of construction land makes it complicated to extract image data. This paper proposes a dimensionality reduction and extraction strategy for the remote sensing data of construction land, with the aid of building information modeling (BIM) and geographical information system (GIS). Firstly, the BIM was employed to reduce the size of the remote sensing data of construction land and to obtain the information of each element. Next, the remote sensing data of construction land were parsed, and the key BIM elements were extracted through semantic filtering. In addition, the remote sensing data were converted into a triangulated irregular network (TIN), which can be processed by the geographical information system (GIS). In the end, random projection was utilized to reduce the dimensionality and compress the remote sensing data, and realize the data extraction. Experimental results show that our approach can compress and extract the information from construction land images in the remote sensing data with a high accuracy.

An enhanced guided LDA model augmented with BERT based semantic strength for aspect term extraction in sentiment analysis

Aspect level sentiment analysis is a fine-grained task in sentiment analysis. It extracts aspects and their corresponding sentiment polarity from opinionated text. The first subtask of identifying the opinionated aspects is called aspect extraction, which is the focus of the work. Social media platforms are an enormous resource of unlabeled data. However, data annotation for fine-grained tasks is quite expensive and laborious. Hence unsupervised models would be highly appreciated. The proposed model is an unsupervised approach for aspect term extraction, a guided Latent Dirichlet Allocation (LDA) model that uses minimal aspect seed words from each aspect category to guide the model in identifying the hidden topics of interest to the user. The guided LDA model is enhanced by guiding inputs using regular expressions based on linguistic rules. The model is further enhanced by multiple pruning strategies, including a BERT based semantic filter, which incorporates semantics to strengthen situations where co-occurrence statistics might fail to serve as a differentiator. The thresholds for these semantic filters have been estimated using Particle Swarm Optimization strategy. The proposed model is expected to overcome the disadvantage of basic LDA models that fail to differentiate the overlapping topics that represent each aspect category. The work has been evaluated on the restaurant domain of SemEval 2014, 2015 and 2016 datasets and has reported an F-measure of 0.81, 0.74 and 0.75 respectively, which is competitive in comparison to the state of art unsupervised baselines and appreciable even with respect to the supervised baselines.

Assessing the Influence Level of Food Safety Public Opinion with Unbalanced Samples Using Ensemble Machine Learning

Assessing the public opinion on food safety events constitutes an important job of government regulators. To optimize the government’s management of food safety affairs, a promising way is to use artificial intelligence to improve the efficiency of food safety public opinion assessment. In this paper, we model the assessment of public opinion influence as a text classification task. The whole model adopts the ensemble learning framework, and it integrates naive Bayes, support vector machine, extreme gradient boosting, convolutional neural network, long- and short-term memory network, FastText, and BERT classification methods into the framework to form an ensemble learner. The ensemble learner is able to classify textual public opinion into high, medium, and low influence levels by learning from the samples assessed by human experts. To overcome the problem of unbalanced samples, we propose a sample generation method consisting of synonym replacement and semantic filtering to increase the number of high-influence samples. Real public opinion data collected from the Food Safety Department of the Chinese government are used for experiment. Extensive comparison of the proposed method with baseline methods proves the effectiveness of the ensemble learner and the sample generation steps.

Real-Time One-Stream Semantic-Guided Refinement Network for RGB-Thermal Salient Object Detection

Salient Object Detection (SOD) has been widely used in practical applications such as multi-sensor image fusion, remote sensing, and defect detection. Recently, SOD from RGB and Thermal (T) has been rapidly developed due to its robustness to extreme situations like low illumination and occlusion. However, existing methods all utilize a dual-stream encoder, which significantly increases the computation burdens and hinders real-world deployment. To this end, we propose a real-time One-stream Semantic-guided Refinement Network (OSRNet) for RGB-T SOD. Specifically, we firstly fuse the RGB and T via concatenation, addition, and multiplication operations to dig the complementary information between each modality. The efficient early fusion not only facilitates the information exchange between each modality but also avoids the cumbersome dual-stream encoder structure. Then, the light-weight decoder is proposed, making the high-level semantic information filter the low-level noisy features and gradually refine the final prediction. Also, we apply deep supervision to make the training procedure more stable and fast. Due to the early fusion strategy, OSRNet can run at a real-time speed (53-60<i>fps</i>) on a single GPU. Extensive quantitative and qualitative experiments show our network outperforms eleven state-of-the-art methods in terms of seven evaluation metrics. Our codes have been released at: https://github.com/huofushuo/OSRNet.

The semantic PHD filter for multi-class target tracking: From theory to practice

In order for a mobile robot to be able to effectively operate in complex, dynamic environments it must be capable of understanding both where and what the objects around them are. In this paper we introduce the semantic probability hypothesis density (SPHD) filter, which allows robots to simultaneously track multiple classes of targets despite measurement uncertainty, including false positive detections, false negative detections, measurement noise, and target misclassification. The SPHD filter is capable of incorporating a different motion model for each type of target and of functioning in situations where the number of targets is unknown and time-varying. To demonstrate the efficacy of the SPHD filter, we conduct both simulated and hardware tests with multiple target types containing both static and dynamic targets. We show that the SPHD filter allows effective tracking of multiple classes of targets even with detection error to some level, and performs better than a collection of PHD filters running in parallel, one for each target class. We also provide a detailed methodology that practitioners can use to fit the probabilistic sensor models necessary to run the SPHD filter.

Hybrid Deep Neural Model for Duplicate Question Detection in Trans-Literated Bi-Lingual Data

Background: Duplicate content often corrupts the filtering mechanism in online question answering. Moreover, as users are usually more comfortable conversing in their native language questions, transliteration adds to the challenges in detecting duplicate questions. This compromises with the response time and increases the answer overload. Thus, it has now become crucial to build clever, intelligent and semantic filters which semantically match linguistically disparate questions. Objective: Most of the research on duplicate question detection has been done on mono-lingual, majorly English Q&amp;A platforms. The aim is to build a model which extends the cognitive capabilities of machines to interpret, comprehend and learn features for semantic matching in transliterated bi-lingual Hinglish (Hindi + English) data acquired from different Q&amp;A platforms. Method: In the proposed DQDHinglish (Duplicate Question Detection) Model, firstly language transformation (transliteration &amp; translation) is done to convert the bi-lingual transliterated question into a monolingual English only text. Next, a hybrid of Siamese neural network containing two identical Long-Term- Short-Memory (LSTM) models and Multi-layer perceptron network is proposed to detect semantically similar question pairs. Manhattan distance function is used as the similarity measure. Result: A dataset was prepared by scrapping 100 question pairs from various social media platforms, such as Quora and TripAdvisor. The performance of the proposed model on the basis of accuracy and Fscore. The proposed DQDHinglish achieves a validation accuracy of 82.40%. Conclusion: A deep neural model was introduced to find semantic match between English question and a Hinglish (Hindi + English) question such that similar intent questions can be combined to enable fast and efficient information processing and delivery. A dataset was created and the proposed model was evaluated on the basis of performance accuracy. To the best of our knowledge, this work is the first reported study on transliterated Hinglish semantic question matching.

Lifelong Learning Courses Recommendation System to Improve Professional Skills Using Ontology and Machine Learning

Lifelong learning enables professionals to update their skills to face challenges in their changing work environments. In view of the wide range of courses on offer, it is important for professionals to have recommendation systems that can link them to suitable courses. Based on this premise and on our previous research, this paper proposes the use of ontology to model job sectors and areas of knowledge, and to represent professional skills that can be automatically updated using the profiled data and machine learning for clustering entities. A three-stage hybrid system is proposed for the recommendation process: semantic filtering, content filtering and heuristics. The proposed system was evaluated with a set of more than 100 user profiles that were used in a previous version of the proposed recommendation system, which allowed the two systems to be compared. The proposed recommender showed 15% improvement when using ontology and clustering with DBSCAN in recall and serendipity metrics, and a six-point increase in harmonic mean over the stored data-based recommender system.

Faster R-CNN Learning-Based Semantic Filter for Geometry Estimation and Its Application in vSLAM Systems

Epipolar geometry is a fundamental constraint used in computer vision systems to estimate parameters using correspondence. The most common way to describe epipolar geometry is by means of a <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$3\times 3$ </tex-math></inline-formula> matrix called the fundamental matrix, and such matrices are used to store the precise geometric information relating a pair of stereo images. Its efficient estimation substantially improves initialization in visual simultaneous localization and mapping (vSLAM), which uses correspondence-based epipolar geometry to determine the trajectory of the camera and a three-dimensional scene. Conventional robust methods for epipolar geometry estimation can become computationally inefficient and inaccurate when there are low-quality correspondences. Because semantic information can be more stable than pixel intensities/descriptors, a novel Faster Region-based Convolutional Network (R-CNN) learning-based approach called the semantic filter is proposed in this paper to address these problems. The semantic filter is first trained on different semantic patches, which are described in terms of their different outlier distributions, providing different semantic labels for image contexts. Then, the patches with low-level semantic labels are filtered out. Finally, precise and robust correspondences can be determined by matches using the high-level semantic contexts, making the correspondence-based calculation more accurate. For dynamic outdoor scenes, the results of extensive experiments show that our semantic filter can help vSLAM localize accurately and robustly on a map from different viewpoints. In a completely static scenario, our semantic filter can remove the low-quality correspondences, enabling the mobile robot to operate well.

An Efficient Module for Instance Segmentation Based on Multi-Level Features and Attention Mechanisms

Recently, multi-level feature networks have been extensively used in instance segmentation. However, because not all features are beneficial to instance segmentation tasks, the performance of networks cannot be adequately improved by synthesizing multi-level convolutional features indiscriminately. In order to solve the problem, an attention-based feature pyramid module (AFPM) is proposed, which integrates the attention mechanism on the basis of a multi-level feature pyramid network to efficiently and pertinently extract the high-level semantic features and low-level spatial structure features; for instance, segmentation. Firstly, we adopt a convolutional block attention module (CBAM) into feature extraction, and sequentially generate attention maps which focus on instance-related features along the channel and spatial dimensions. Secondly, we build inter-dimensional dependencies through a convolutional triplet attention module (CTAM) in lateral attention connections, which is used to propagate a helpful semantic feature map and filter redundant informative features irrelevant to instance objects. Finally, we construct branches for feature enhancement to strengthen detailed information to boost the entire feature hierarchy of the network. The experimental results on the Cityscapes dataset manifest that the proposed module outperforms other excellent methods under different evaluation metrics and effectively upgrades the performance of the instance segmentation method.

Thematic category of meal, cuisine, and hospitality in the Hungarian proverbial lore

This study is aimed on the thematic selection of the Hungarian proverbs dealing with gastronomy and hospitality both in direct and figurative way. As a source for this compilation we used recent and earlier editions of the proverbs’ dictionaries. We searched for proverbs using keywords, and not only. A substantial number of proverbs was identified as belonging to this thematic group. Through the comparative and descriptive approach, we divided all the selected proverbial material into subgroups. A special semantic filter was applied to distinguish the proverbs containing so-called Hungarikum realia. The selection of Hungarian proverbs for this thematic groups was also used in order to contribute to the theory of the common Central-European linguistic area.