A Framework for Automated Scraping of Structured Data Records From the Deep Web Using Semantic Labeling

The intent of this research is to come up with an automated web scraping system which is capable of extracting structured data records embedded in semi-structured web pages. Most of the automated extraction techniques in the literature captures repeated pattern among a set of similarly structured web pages, thereby deducing the template used for the generation of those web pages and then data records extraction is done. All of these techniques exploit computationally intensive operations such as string pattern matching or DOM tree matching and then perform manual labeling of extracted data records. The technique discussed in this paper departs from the state-of-the-art approaches by determining informative sections in the web page through repetition of informative content rather than syntactic structure. From the experiments, it is clear that the system has identified data rich region with 100% precision for web sites belonging to different domains. The experiments conducted on the real world web sites prove the effectiveness and versatility of the proposed approach.

Lane following Learning Based on Semantic Segmentation with Chroma Key and Image Superposition

There are various techniques to approach learning in autonomous driving; however, all of them suffer from some problems. In the case of imitation learning based on artificial neural networks, the system must learn to correctly identify the elements of the environment. In some cases, it takes a lot of effort to tag the images with the proper semantics. This is also relevant given the need to have very varied scenarios to train and to thus obtain an acceptable generalization capacity. In the present work, we propose a technique for automated semantic labeling. It is based on various learning phases using image superposition combining both scenarios with chromas and real indoor scenarios. This allows the generation of augmented datasets that facilitate the learning process. Further improvements by applying noise techniques are also studied. To carry out the validation, a small-scale car model is used that learns to automatically drive on a reduced circuit. A comparison with models that do not rely on semantic segmentation is also performed. The main contribution of our proposal is the possibility of generating datasets for real indoor scenarios with automatic semantic segmentation, without the need for endless human labeling tasks.

Design and Implementation of an Ontology for Semantic Labeling and Testing: Automotive Global Ontology (AGO)

Modern Artificial Intelligence (AI) methods can produce a large quantity of accurate and richly described data, in domains such as surveillance or automation. As a result, the need to organize data at a large scale in a semantic structure has arisen for long-term data maintenance and consumption. Ontologies and graph databases have gained popularity as mechanisms to satisfy this need. Ontologies provide the means to formally structure descriptive and semantic relations of a domain. Graph databases allow efficient and well-adapted storage, manipulation, and consumption of these linked data resources. However, at present, there is no a universally defined strategy for building AI-oriented ontologies for the automotive sector. One of the key challenges is the lack of a global standardized vocabulary. Most private initiatives and large open datasets for Advanced Driver Assistance Systems (ADASs) and Autonomous Driving (AD) development include their own definitions of terms, with incompatible taxonomies and structures, thus resulting in a well-known lack of interoperability. This paper presents the Automotive Global Ontology (AGO) as a Knowledge Organization System (KOS) using a graph database (Neo4j). Two different use cases for the AGO domain ontology are presented to showcase its capabilities in terms of semantic labeling and scenario-based testing. The ontology and related material have been made public for their subsequent use by the industry and academic communities.

Class-Wise Fully Convolutional Network for Semantic Segmentation of Remote Sensing Images

Semantic segmentation is a fundamental task in remote sensing image interpretation, which aims to assign a semantic label for every pixel in the given image. Accurate semantic segmentation is still challenging due to the complex distributions of various ground objects. With the development of deep learning, a series of segmentation networks represented by fully convolutional network (FCN) has made remarkable progress on this problem, but the segmentation accuracy is still far from expectations. This paper focuses on the importance of class-specific features of different land cover objects, and presents a novel end-to-end class-wise processing framework for segmentation. The proposed class-wise FCN (C-FCN) is shaped in the form of an encoder-decoder structure with skip-connections, in which the encoder is shared to produce general features for all categories and the decoder is class-wise to process class-specific features. To be detailed, class-wise transition (CT), class-wise up-sampling (CU), class-wise supervision (CS), and class-wise classification (CC) modules are designed to achieve the class-wise transfer, recover the resolution of class-wise feature maps, bridge the encoder and modified decoder, and implement class-wise classifications, respectively. Class-wise and group convolutions are adopted in the architecture with regard to the control of parameter numbers. The method is tested on the public ISPRS 2D semantic labeling benchmark datasets. Experimental results show that the proposed C-FCN significantly improves the segmentation performances compared with many state-of-the-art FCN-based networks, revealing its potentials on accurate segmentation of complex remote sensing images.

Artificial intelligence in CT for quantifying lung changes in the era of CFTR modulators

RationaleChest computed tomography (CT) remains the imaging standard for demonstrating cystic fibrosis airway structural disease in vivo. However, visual scorings as an outcome measure are time-consuming, require training, and lack high reproducibility.ObjectiveTo validate a fully automated artificial intelligence-driven scoring of cystic fibrosis lung disease severity.MethodsData were retrospectively collected in three cystic fibrosis reference centers, between 2008 and 2020, in 184 patients 4 to 54-years-old. An algorithm using three two-dimensional convolutional neural networks was trained with 78 patients’ CTs (23 530 CT slices) for the semantic labeling of bronchiectasis, peribronchial thickening, bronchial mucus, bronchiolar mucus, and collapse/consolidation. 36 patients’ CTs (11 435 CT slices) were used for testing versus ground-truth labels. The method's clinical validity was assessed in an independent group of 70 patients with or without lumacaftor/ivacaftor treatment (n=10 and 60, respectively) with repeat examinations. Similarity and reproducibility were assessed using Dice coefficient, correlations using Spearman test, and paired comparisons using Wilcoxon rank test.Measurement and main resultsThe overall pixelwise similarity of artificial intelligence-driven versus ground-truth labels was good (Dice coefficient=0.71). All artificial intelligence-driven volumetric quantifications had moderate to very good correlations to a visual imaging scoring (p<0.001) and fair to good correlations to FEV1% at pulmonary function test (p<0.001). Significant decreases in peribronchial thickening (p=0.005), bronchial mucus (p=0.005), bronchiolar mucus (p=0.007) volumes were measured in patients with lumacaftor/ivacaftor. Conversely, bronchiectasis (p=0.002) and peribronchial thickening (p=0.008) volumes increased in patients without lumacaftor/ivacaftor. The reproducibility was almost perfect (Dice>0.99).ConclusionArtificial intelligence allows a fully automated volumetric quantification of cystic fibrosis-related modifications over an entire lung. The novel scoring system could provide a robust disease outcome in the era of effective CFTR modulator therapy.

MARRYING DEEP LEARNING AND DATA FUSION FOR ACCURATE SEMANTIC LABELING OF SENTINEL-2 IMAGES

The understanding of the Earth through global land monitoring from satellite images paves the way towards many applications including flight simulations, urban management and telecommunications. The twin satellites from the Sentinel-2 mission developed by the European Space Agency (ESA) provide 13 spectral bands with a high observation frequency worldwide. In this paper, we present a novel multi-temporal approach for land-cover classification of Sentinel-2 images whereby a time-series of images is classified using fully convolutional network U-Net models and then coupled by a developed probabilistic algorithm. The proposed pipeline further includes an automatic quality control and correction step whereby an external source can be introduced in order to validate and correct the deep learning classification. The final step consists of adjusting the combined predictions to the cloud-free mosaic built from Sentinel-2 L2A images in order for the classification to more closely match the reference mosaic image.