Superpixel-Based Attention Graph Neural Network for Semantic Segmentation in Aerial Images

Semantic segmentation is one of the significant tasks in understanding aerial images with high spatial resolution. Recently, Graph Neural Network (GNN) and attention mechanism have achieved excellent performance in semantic segmentation tasks in general images and been applied to aerial images. In this paper, we propose a novel Superpixel-based Attention Graph Neural Network (SAGNN) for semantic segmentation of high spatial resolution aerial images. A K-Nearest Neighbor (KNN) graph is constructed from our network for each image, where each node corresponds to a superpixel in the image and is associated with a hidden representation vector. On this basis, the initialization of the hidden representation vector is the appearance feature extracted by a unary Convolutional Neural Network (CNN) from the image. Moreover, relying on the attention mechanism and recursive functions, each node can update its hidden representation according to the current state and the incoming information from its neighbors. The final representation of each node is used to predict the semantic class of each superpixel. The attention mechanism enables graph nodes to differentially aggregate neighbor information, which can extract higher-quality features. Furthermore, the superpixels not only save computational resources, but also maintain object boundary to achieve more accurate predictions. The accuracy of our model on the Potsdam and Vaihingen public datasets exceeds all benchmark approaches, reaching 90.23% and 89.32%, respectively.

‘Path’ and ‘Manner’ in The Serialization of Motion Verbs

Verbs are one of the most basic vocabulary elements of the language. These words express the being, manners and movements of beings in the universe. However, in order to express the movements of living and non-living beings in the universe, the verb category creates syntactic and semantic situations by performing some combinations within itself. Because the many movements of many beings in the universe cause endless combinations to appear. In order to express this, the verb category creates various combinations within itself. The most typical example of this is seen between motion verbs, which is a semantic class, and serial verb structures, which are a syntactic and semantic class. Although there have been various studies on motion verbs, the limits of these verbs have not been determined in Turkish studies. Motion verbs, which can be considered as verbs expressing the displacement of beings in the universe, are closely related to serial verb structures, which is a syntactic and semantic category. Serial verb structures contain at least one motion verb in surface or deep structure. According to Talmy's typology, these verbs of motion mark the "way" where the movement takes place, or the "style", which is the way it takes place. In this study, the roles of "path" and "manner" in the serialization process of motion verbs are discussed in the Turkish context.

Demonstratiivpronoomenid ja -adverbid määratlejatena. Miks me oleme siin ilmas, selles olukorras?

Kokkuvõte. Artiklis analüüsime eesti keele demonstratiivide referentsiaalseid omadusi sellistes konstruktsioonides, kus demonstratiivid kuuluvad definiitse määratlejana nimisõnafraasi koosseisu. Otsime vastust küsimusele, mille poolest erinevad demonstratiivadverb (nt siin, seal) ning demonstratiivpronoomen (see, too), kui need esinevad määratlejana koos kohakäändes nimisõnafraasiga (vrd siin koolis ja selles koolis). Oleme püstitanud hüpoteesi, et demonstratiivadverbid seostuvad ruumitähendust väljendavate substantiividega, demonstratiivpronoomenid esinevad aga nende substantiividega, mille referent on mitteruumiline. Uurimuse andmestik pärineb 2017. aasta eesti keele ühendkorpusest, kust oleme võtnud 100 lauset iga demonstratiivi kohta igas kohakäändes, seega kokku 2400 lauset. Materjali analüüsime kvantitatiivselt (tingimuslike otsustuspuude ja juhumetsadega) ning kvalitatiivselt. Uurimuse tulemused kinnitavad, et substantiivi semantilised omadused, täpsemalt substantiivi semantiline klass ning konkreetsus, on seotud määratleja valikuga. Kohatähenduses substantiividega esineb määratlejana sagedamini demonstratiivadverb, mittekoha tähenduses substantiivide määratlejana kasutatakse aga demonstratiivpronoomenit. Mittekohta tähistavate substantiivide korral mõjutab määratleja valikut omakorda sõna konkreetsus. Seega on võimalik demonstratiivseid määratlejaid eesti keeles kasutada referenti looval viisil. Abstract. Helen Hint, Piia Taremaa, Maria Reile, Renate Pajusalu: Demonstrative pronouns and demonstrative adverbs as determiners in Estonian: why are we in “here world” in “this situation”? We investigate the variation of definite determiner constructions in Estonian: noun phrases with a demonstrative pronoun (see ‘this’, too ‘that’) or demonstrative adverb (siin ‘here’, seal ‘there’) as a determiner are contrasted. The question is what differentiates the use of a demonstrative pronoun and a demonstrative adverb if used in a determiner position in an NP. The data from Estonian National Corpus 2017 were tagged for semantic class of a noun, noun concreteness, and verb type. We collected 100 clauses for each sub-construction (six spatial cases crossed with four determiner forms), 2400 clauses in total. For statistical analysis, we used conditional random forests and inference trees. We show that nouns expressing spatial meaning prefer demonstrative adverbs as determiners, while non-spatial nouns combine with demonstrative pronouns. Spatiality-wise polysemous nouns exhibit more varied preferences. Adverbial determiners are more probably used with concrete nouns, and abstract nouns co-occur with pronominals. Overall, the frequency of demonstrative adverbs as NP attributes confirms that demonstrative adverbs are productive determiners in Estonian.

Interface Delay

Interface Delay is a theory of syntactic development, which attempts to explain an array of constructions that are slow to develop, which are characterized by being sensitive to discourse-pragmatic considerations of the type associated with the natural semantic class of definites. The theory claims that neither syntax itself, nor the discourse-pragmatic abilities related to executive function and theory of mind themselves are slow to develop. Rather, the claim is that the nexus or interface between the two cognitive domains is slow to develop. We review the development of subjects in child Spanish as an example of this delayed growth trajectory. Further, we review evidence that a delay in the development of tense causes concomitant delays in the seemingly unrelated phenomena of non-nominative case subject pronoun use and un-inverted wh- questions.

Translation Approaches in Rendering Names of Tourist Sites

This article identifies the translation approaches adopted in the translation of names of tourist sites in China and examines how ‘fame’ and ‘popularity’ may influence these approaches. Upon analyzing a corpus of scenic site names, it is found that ‘pure phonetic’, ‘phonetic (name) + semantic (class)’, ‘pure semantic’, and ‘phonetic (location) + semantic (name) are the four major patterns in the translations of site names. On the whole, the data shows that phonetic translation is dominant over semantic translation. Meanwhile, ‘fame’ and ‘popularity’ have great impact on the translated names of scenic sites. The findings also suggest that a phonetic translation approach is preferred in rendering names of world-famous sites whereas a semantic translation approach is more frequently used for the name translation of sites located in places with higher popularity. The conflicting results reflect China’s struggle between preserving its cultural flavor for the sake of national identity and catering to foreign visitors for the benefit of the country’s tourism development.

Region-Based Static Video Stitching for Reduction of Parallax Distortion

In this paper, we propose a semantic segmentation-based static video stitching method to reduce parallax and misalignment distortion for sports stadium scenes with dynamic foreground objects. First, video frame pairs for stitching are divided into segments of different classes through semantic segmentation. Region-based stitching is performed on matched segment pairs, assuming that segments of the same semantic class are on the same plane. Second, to prevent degradation of the stitching quality of plain or noisy videos, the homography for each matched segment pair is estimated using the temporally consistent feature points. Finally, the stitched video frame is synthesized by stacking the stitched matched segment pairs and the foreground segments to the reference frame plane by descending order of the area. The performance of the proposed method is evaluated by comparing the subjective quality, geometric distortion, and pixel distortion of video sequences stitched using the proposed and conventional methods. The proposed method is shown to reduce parallax and misalignment distortion in segments with plain texture or large parallax, and significantly improve geometric distortion and pixel distortion compared to conventional methods.

Multi-Scale Squeeze U-SegNet with Multi Global Attention for Brain MRI Segmentation

In this paper, we propose a multi-scale feature extraction with novel attention-based convolutional learning using the U-SegNet architecture to achieve segmentation of brain tissue from a magnetic resonance image (MRI). Although convolutional neural networks (CNNs) show enormous growth in medical image segmentation, there are some drawbacks with the conventional CNN models. In particular, the conventional use of encoder-decoder approaches leads to the extraction of similar low-level features multiple times, causing redundant use of information. Moreover, due to inefficient modeling of long-range dependencies, each semantic class is likely to be associated with non-accurate discriminative feature representations, resulting in low accuracy of segmentation. The proposed global attention module refines the feature extraction and improves the representational power of the convolutional neural network. Moreover, the attention-based multi-scale fusion strategy can integrate local features with their corresponding global dependencies. The integration of fire modules in both the encoder and decoder paths can significantly reduce the computational complexity owing to fewer model parameters. The proposed method was evaluated on publicly accessible datasets for brain tissue segmentation. The experimental results show that our proposed model achieves segmentation accuracies of 94.81% for cerebrospinal fluid (CSF), 95.54% for gray matter (GM), and 96.33% for white matter (WM) with a noticeably reduced number of learnable parameters. Our study shows better segmentation performance, improving the prediction accuracy by 2.5% in terms of dice similarity index while achieving a 4.5 times reduction in the number of learnable parameters compared to previously developed U-SegNet based segmentation approaches. This demonstrates that the proposed approach can achieve reliable and precise automatic segmentation of brain MRI images.

The role of verb semantics in Hungarian verb-object order

Hungarian is often referred to as a discourse-configurational language, since the structural position of constituents is determined by their logical function (topic or comment) rather than their grammatical function (e.g., subject or object). We build on work by Komlósy (1989) and argue that in addition to discourse context, the lexical semantics of the verb also plays a significant role in determining Hungarian word order. In order to investigate the role of lexical semantics in determining Hungarian word order, we conduct a large-scale, data-driven analysis on the ordering of 380 transitive verbs and their objects, as observed in hundreds of thousands of examples extracted from the Hungarian Gigaword Corpus. We test the effect of lexical semantics on the ordering of verbs and their objects by grouping verbs into 11 semantic classes. In addition to the semantic class of the verb, we also include two control features related to information structure, object definiteness and object NP weight, chosen to allow a comparison of their effect size to that of verb semantics. Our results suggest that all three features have a significant effect on verb-object ordering in Hungarian and among these features, the semantic class of the verb has the largest effect. Specifically, we find that stative verbs, such as fed 'cover', jelent 'mean' and övez 'surround', tend to be OV-preferring (with the exception of psych verbs which are strongly VO-preferring) and non-stative verbs, such as bírál 'judge', csökkent 'reduce' and csókol 'kiss', verbs tend to be VO-preferring. These findings support our hypothesis that lexical semantic factors influence word order in Hungarian.

Semantically Derived Geometric Constraints for MVS Reconstruction of Textureless Areas

Conventional multi-view stereo (MVS) approaches based on photo-consistency measures are generally robust, yet often fail in calculating valid depth pixel estimates in low textured areas of the scene. In this study, a novel approach is proposed to tackle this challenge by leveraging semantic priors into a PatchMatch-based MVS in order to increase confidence and support depth and normal map estimation. Semantic class labels on image pixels are used to impose class-specific geometric constraints during multiview stereo, optimising the depth estimation on weakly supported, textureless areas, commonly present in urban scenarios of building facades, indoor scenes, or aerial datasets. Detecting dominant shapes, e.g., planes, with RANSAC, an adjusted cost function is introduced that combines and weighs both photometric and semantic scores propagating, thus, more accurate depth estimates. Being adaptive, it fills in apparent information gaps and smoothing local roughness in problematic regions while at the same time preserves important details. Experiments on benchmark and custom datasets demonstrate the effectiveness of the presented approach.

TO THE STUDY OF THE SEMANTICS OF THE VERB

In this article, verbal lexemes are classified according to the nomination of the activity of nouns. Consequently, they are called upon to denote what is the “characteristic activity” of nouns of specific semantic classes, semantic thematic series and individual lexemes, as well as to the principle of generalization of different semantic classes, a group of thematic series, which is proved on specific verb examples and it is necessary to conclude that verbs can be divided into verbs of narrow and wide nominations, Depending on the semantic structure of verbs, their direct and figurative meanings differ, Often the potential seme of a verb is a concretizer and indicates the semantic class, groups and thematic series of nouns, and thus the verb actualizes its meaning in speech.The starting point of this work is the fact that "there are no objects without properties and relations and properties and relations without objects", therefore, verbs as well as nouns can be subjected to such classifications as nouns, where nouns of being, abstractness, concreteness, animate, inanimate are distinguished, anthroponymy, faunonymy, as well as certain semantic groups, thematic series and at the level of individual lexemes, as indicated by specific examples.Thus, we have to conclude that the verb is designed in the language to designate the characteristic activity of certain nouns, combining with it in speech its actual meaning is revealed and thereby determines its relevance to a particular semantic class, semantic groups or thematic series, and thus the verbs of a narrow and a wide nomination from a wide nomination. Depending on the semantic structure of the verbs, their direct and figurative meanings differ. Often a potential seme of a verb is a concretizer and indicates the semantic class, groups and thematic series of nouns, and thus the verb actualizes its meaning in speech.