Search results for: text detection
Commenced in January 2007
Frequency: Monthly
Edition: International
Paper Count: 4513

Search results for: text detection

4513 On-Road Text Detection Platform for Driver Assistance Systems

Authors: Guezouli Larbi, Belkacem Soundes

Abstract:

The automation of the text detection process can help the human in his driving task. Its application can be very useful to help drivers to have more information about their environment by facilitating the reading of road signs such as directional signs, events, stores, etc. In this paper, a system consisting of two stages has been proposed. In the first one, we used pseudo-Zernike moments to pinpoint areas of the image that may contain text. The architecture of this part is based on three main steps, region of interest (ROI) detection, text localization, and non-text region filtering. Then, in the second step, we present a convolutional neural network architecture (On-Road Text Detection Network - ORTDN) which is considered a classification phase. The results show that the proposed framework achieved ≈ 35 fps and an mAP of ≈ 90%, thus a low computational time with competitive accuracy.

Keywords: text detection, CNN, PZM, deep learning

Procedia PDF Downloads 45
4512 Anatomical Survey for Text Pattern Detection

Authors: S. Tehsin, S. Kausar

Abstract:

The ultimate aim of machine intelligence is to explore and materialize the human capabilities, one of which is the ability to detect various text objects within one or more images displayed on any canvas including prints, videos or electronic displays. Multimedia data has increased rapidly in past years. Textual information present in multimedia contains important information about the image/video content. However, it needs to technologically testify the commonly used human intelligence of detecting and differentiating the text within an image, for computers. Hence in this paper feature set based on anatomical study of human text detection system is proposed. Subsequent examination bears testimony to the fact that the features extracted proved instrumental to text detection.

Keywords: biologically inspired vision, content based retrieval, document analysis, text extraction

Procedia PDF Downloads 412
4511 Video Text Information Detection and Localization in Lecture Videos Using Moments

Authors: Belkacem Soundes, Guezouli Larbi

Abstract:

This paper presents a robust and accurate method for text detection and localization over lecture videos. Frame regions are classified into text or background based on visual feature analysis. However, lecture video shows significant degradation mainly related to acquisition conditions, camera motion and environmental changes resulting in low quality videos. Hence, affecting feature extraction and description efficiency. Moreover, traditional text detection methods cannot be directly applied to lecture videos. Therefore, robust feature extraction methods dedicated to this specific video genre are required for robust and accurate text detection and extraction. Method consists of a three-step process: Slide region detection and segmentation; Feature extraction and non-text filtering. For robust and effective features extraction moment functions are used. Two distinct types of moments are used: orthogonal and non-orthogonal. For orthogonal Zernike Moments, both Pseudo Zernike moments are used, whereas for non-orthogonal ones Hu moments are used. Expressivity and description efficiency are given and discussed. Proposed approach shows that in general, orthogonal moments show high accuracy in comparison to the non-orthogonal one. Pseudo Zernike moments are more effective than Zernike with better computation time.

Keywords: text detection, text localization, lecture videos, pseudo zernike moments

Procedia PDF Downloads 114
4510 Extraction of Text Subtitles in Multimedia Systems

Authors: Amarjit Singh

Abstract:

In this paper, a method for extraction of text subtitles in large video is proposed. The video data needs to be annotated for many multimedia applications. Text is incorporated in digital video for the motive of providing useful information about that video. So need arises to detect text present in video to understanding and video indexing. This is achieved in two steps. First step is text localization and the second step is text verification. The method of text detection can be extended to text recognition which finds applications in automatic video indexing; video annotation and content based video retrieval. The method has been tested on various types of videos.

Keywords: video, subtitles, extraction, annotation, frames

Procedia PDF Downloads 557
4509 Distorted Document Images Dataset for Text Detection and Recognition

Authors: Ilia Zharikov, Philipp Nikitin, Ilia Vasiliev, Vladimir Dokholyan

Abstract:

With the increasing popularity of document analysis and recognition systems, text detection (TD) and optical character recognition (OCR) in document images become challenging tasks. However, according to our best knowledge, no publicly available datasets for these particular problems exist. In this paper, we introduce a Distorted Document Images dataset (DDI-100) and provide a detailed analysis of the DDI-100 in its current state. To create the dataset we collected 7000 unique document pages, and extend it by applying different types of distortions and geometric transformations. In total, DDI-100 contains more than 100,000 document images together with binary text masks, text and character locations in terms of bounding boxes. We also present an analysis of several state-of-the-art TD and OCR approaches on the presented dataset. Lastly, we demonstrate the usefulness of DDI-100 to improve accuracy and stability of the considered TD and OCR models.

Keywords: document analysis, open dataset, optical character recognition, text detection

Procedia PDF Downloads 128
4508 Urdu Text Extraction Method from Images

Authors: Samabia Tehsin, Sumaira Kausar

Abstract:

Due to the vast increase in the multimedia data in recent years, efficient and robust retrieval techniques are needed to retrieve and index images/ videos. Text embedded in the images can serve as the strong retrieval tool for images. This is the reason that text extraction is an area of research with increasing attention. English text extraction is the focus of many researchers but very less work has been done on other languages like Urdu. This paper is focusing on Urdu text extraction from video frames. This paper presents a text detection feature set, which has the ability to deal up with most of the problems connected with the text extraction process. To test the validity of the method, it is tested on Urdu news dataset, which gives promising results.

Keywords: caption text, content-based image retrieval, document analysis, text extraction

Procedia PDF Downloads 472
4507 Off-Topic Text Detection System Using a Hybrid Model

Authors: Usama Shahid

Abstract:

Be it written documents, news columns, or students' essays, verifying the content can be a time-consuming task. Apart from the spelling and grammar mistakes, the proofreader is also supposed to verify whether the content included in the essay or document is relevant or not. The irrelevant content in any document or essay is referred to as off-topic text and in this paper, we will address the problem of off-topic text detection from a document using machine learning techniques. Our study aims to identify the off-topic content from a document using Echo state network model and we will also compare data with other models. The previous study uses Convolutional Neural Networks and TFIDF to detect off-topic text. We will rearrange the existing datasets and take new classifiers along with new word embeddings and implement them on existing and new datasets in order to compare the results with the previously existing CNN model.

Keywords: off topic, text detection, eco state network, machine learning

Procedia PDF Downloads 42
4506 A Character Detection Method for Ancient Yi Books Based on Connected Components and Regressive Character Segmentation

Authors: Xu Han, Shanxiong Chen, Shiyu Zhu, Xiaoyu Lin, Fujia Zhao, Dingwang Wang

Abstract:

Character detection is an important issue for character recognition of ancient Yi books. The accuracy of detection directly affects the recognition effect of ancient Yi books. Considering the complex layout, the lack of standard typesetting and the mixed arrangement between images and texts, we propose a character detection method for ancient Yi books based on connected components and regressive character segmentation. First, the scanned images of ancient Yi books are preprocessed with nonlocal mean filtering, and then a modified local adaptive threshold binarization algorithm is used to obtain the binary images to segment the foreground and background for the images. Second, the non-text areas are removed by the method based on connected components. Finally, the single character in the ancient Yi books is segmented by our method. The experimental results show that the method can effectively separate the text areas and non-text areas for ancient Yi books and achieve higher accuracy and recall rate in the experiment of character detection, and effectively solve the problem of character detection and segmentation in character recognition of ancient books.

Keywords: CCS concepts, computing methodologies, interest point, salient region detections, image segmentation

Procedia PDF Downloads 90
4505 Adaptation of Projection Profile Algorithm for Skewed Handwritten Text Line Detection

Authors: Kayode A. Olaniyi, Tola. M. Osifeko, Adeola A. Ogunleye

Abstract:

Text line segmentation is an important step in document image processing. It represents a labeling process that assigns the same label using distance metric probability to spatially aligned units. Text line detection techniques have successfully been implemented mainly in printed documents. However, processing of the handwritten texts especially unconstrained documents has remained a key problem. This is because the unconstrained hand-written text lines are often not uniformly skewed. The spaces between text lines may not be obvious, complicated by the nature of handwriting and, overlapping ascenders and/or descenders of some characters. Hence, text lines detection and segmentation represents a leading challenge in handwritten document image processing. Text line detection methods that rely on the traditional global projection profile of the text document cannot efficiently confront with the problem of variable skew angles between different text lines. Hence, the formulation of a horizontal line as a separator is often not efficient. This paper presents a technique to segment a handwritten document into distinct lines of text. The proposed algorithm starts, by partitioning the initial text image into columns, across its width into chunks of about 5% each. At each vertical strip of 5%, the histogram of horizontal runs is projected. We have worked with the assumption that text appearing in a single strip is almost parallel to each other. The algorithm developed provides a sliding window through the first vertical strip on the left side of the page. It runs through to identify the new minimum corresponding to a valley in the projection profile. Each valley would represent the starting point of the orientation line and the ending point is the minimum point on the projection profile of the next vertical strip. The derived text-lines traverse around any obstructing handwritten vertical strips of connected component by associating it to either the line above or below. A decision of associating such connected component is made by the probability obtained from a distance metric decision. The technique outperforms the global projection profile for text line segmentation and it is robust to handle skewed documents and those with lines running into each other.

Keywords: connected-component, projection-profile, segmentation, text-line

Procedia PDF Downloads 82
4504 Fake News Detection for Korean News Using Machine Learning Techniques

Authors: Tae-Uk Yun, Pullip Chung, Kee-Young Kwahk, Hyunchul Ahn

Abstract:

Fake news is defined as the news articles that are intentionally and verifiably false, and could mislead readers. Spread of fake news may provoke anxiety, chaos, fear, or irrational decisions of the public. Thus, detecting fake news and preventing its spread has become very important issue in our society. However, due to the huge amount of fake news produced every day, it is almost impossible to identify it by a human. Under this context, researchers have tried to develop automated fake news detection using machine learning techniques over the past years. But, there have been no prior studies proposed an automated fake news detection method for Korean news to our best knowledge. In this study, we aim to detect Korean fake news using text mining and machine learning techniques. Our proposed method consists of two steps. In the first step, the news contents to be analyzed is convert to quantified values using various text mining techniques (topic modeling, TF-IDF, and so on). After that, in step 2, classifiers are trained using the values produced in step 1. As the classifiers, machine learning techniques such as logistic regression, backpropagation network, support vector machine, and deep neural network can be applied. To validate the effectiveness of the proposed method, we collected about 200 short Korean news from Seoul National University’s FactCheck. which provides with detailed analysis reports from 20 media outlets and links to source documents for each case. Using this dataset, we will identify which text features are important as well as which classifiers are effective in detecting Korean fake news.

Keywords: fake news detection, Korean news, machine learning, text mining

Procedia PDF Downloads 234
4503 A Summary-Based Text Classification Model for Graph Attention Networks

Authors: Shuo Liu

Abstract:

In Chinese text classification tasks, redundant words and phrases can interfere with the formation of extracted and analyzed text information, leading to a decrease in the accuracy of the classification model. To reduce irrelevant elements, extract and utilize text content information more efficiently and improve the accuracy of text classification models. In this paper, the text in the corpus is first extracted using the TextRank algorithm for abstraction, the words in the abstract are used as nodes to construct a text graph, and then the graph attention network (GAT) is used to complete the task of classifying the text. Testing on a Chinese dataset from the network, the classification accuracy was improved over the direct method of generating graph structures using text.

Keywords: Chinese natural language processing, text classification, abstract extraction, graph attention network

Procedia PDF Downloads 56
4502 Development of Fake News Model Using Machine Learning through Natural Language Processing

Authors: Sajjad Ahmed, Knut Hinkelmann, Flavio Corradini

Abstract:

Fake news detection research is still in the early stage as this is a relatively new phenomenon in the interest raised by society. Machine learning helps to solve complex problems and to build AI systems nowadays and especially in those cases where we have tacit knowledge or the knowledge that is not known. We used machine learning algorithms and for identification of fake news; we applied three classifiers; Passive Aggressive, Naïve Bayes, and Support Vector Machine. Simple classification is not completely correct in fake news detection because classification methods are not specialized for fake news. With the integration of machine learning and text-based processing, we can detect fake news and build classifiers that can classify the news data. Text classification mainly focuses on extracting various features of text and after that incorporating those features into classification. The big challenge in this area is the lack of an efficient way to differentiate between fake and non-fake due to the unavailability of corpora. We applied three different machine learning classifiers on two publicly available datasets. Experimental analysis based on the existing dataset indicates a very encouraging and improved performance.

Keywords: fake news detection, natural language processing, machine learning, classification techniques.

Procedia PDF Downloads 119
4501 Small Text Extraction from Documents and Chart Images

Authors: Rominkumar Busa, Shahira K. C., Lijiya A.

Abstract:

Text recognition is an important area in computer vision which deals with detecting and recognising text from an image. The Optical Character Recognition (OCR) is a saturated area these days and with very good text recognition accuracy. However the same OCR methods when applied on text with small font sizes like the text data of chart images, the recognition rate is less than 30%. In this work, aims to extract small text in images using the deep learning model, CRNN with CTC loss. The text recognition accuracy is found to improve by applying image enhancement by super resolution prior to CRNN model. We also observe the text recognition rate further increases by 18% by applying the proposed method, which involves super resolution and character segmentation followed by CRNN with CTC loss. The efficiency of the proposed method shows that further pre-processing on chart image text and other small text images will improve the accuracy further, thereby helping text extraction from chart images.

Keywords: small text extraction, OCR, scene text recognition, CRNN

Procedia PDF Downloads 82
4500 Text Data Preprocessing Library: Bilingual Approach

Authors: Kabil Boukhari

Abstract:

In the context of information retrieval, the selection of the most relevant words is a very important step. In fact, the text cleaning allows keeping only the most representative words for a better use. In this paper, we propose a library for the purpose text preprocessing within an implemented application to facilitate this task. This study has two purposes. The first, is to present the related work of the various steps involved in text preprocessing, presenting the segmentation, stemming and lemmatization algorithms that could be efficient in the rest of study. The second, is to implement a developed tool for text preprocessing in French and English. This library accepts unstructured text as input and provides the preprocessed text as output, based on a set of rules and on a base of stop words for both languages. The proposed library has been made on different corpora and gave an interesting result.

Keywords: text preprocessing, segmentation, knowledge extraction, normalization, text generation, information retrieval

Procedia PDF Downloads 50
4499 Spontaneous Message Detection of Annoying Situation in Community Networks Using Mining Algorithm

Authors: P. Senthil Kumari

Abstract:

Main concerns in data mining investigation are social controls of data mining for handling ambiguity, noise, or incompleteness on text data. We describe an innovative approach for unplanned text data detection of community networks achieved by classification mechanism. In a tangible domain claim with humble secrecy backgrounds provided by community network for evading annoying content is presented on consumer message partition. To avoid this, mining methodology provides the capability to unswervingly switch the messages and similarly recover the superiority of ordering. Here we designated learning-centered mining approaches with pre-processing technique to complete this effort. Our involvement of work compact with rule-based personalization for automatic text categorization which was appropriate in many dissimilar frameworks and offers tolerance value for permits the background of comments conferring to a variety of conditions associated with the policy or rule arrangements processed by learning algorithm. Remarkably, we find that the choice of classifier has predicted the class labels for control of the inadequate documents on community network with great value of effect.

Keywords: text mining, data classification, community network, learning algorithm

Procedia PDF Downloads 467
4498 Literature Review on Text Comparison Techniques: Analysis of Text Extraction, Main Comparison and Visual Representation Tools

Authors: Andriana Mkrtchyan, Vahe Khlghatyan

Abstract:

The choice of a profession is one of the most important decisions people make throughout their life. With the development of modern science, technologies, and all the spheres existing in the modern world, more and more professions are being arisen that complicate even more the process of choosing. Hence, there is a need for a guiding platform to help people to choose a profession and the right career path based on their interests, skills, and personality. This review aims at analyzing existing methods of comparing PDF format documents and suggests that a 3-stage approach is implemented for the comparison, that is – 1. text extraction from PDF format documents, 2. comparison of the extracted text via NLP algorithms, 3. comparison representation using special shape and color psychology methodology.

Keywords: color psychology, data acquisition/extraction, data augmentation, disambiguation, natural language processing, outlier detection, semantic similarity, text-mining, user evaluation, visual search

Procedia PDF Downloads 26
4497 Binarization and Recognition of Characters from Historical Degraded Documents

Authors: Bency Jacob, S.B. Waykar

Abstract:

Degradations in historical document images appear due to aging of the documents. It is very difficult to understand and retrieve text from badly degraded documents as there is variation between the document foreground and background. Thresholding of such document images either result in broken characters or detection of false texts. Numerous algorithms exist that can separate text and background efficiently in the textual regions of the document; but portions of background are mistaken as text in areas that hardly contain any text. This paper presents a way to overcome these problems by a robust binarization technique that recovers the text from a severely degraded document images and thereby increases the accuracy of optical character recognition systems. The proposed document recovery algorithm efficiently removes degradations from document images. Here we are using the ostus method ,local thresholding and global thresholding and after the binarization training and recognizing the characters in the degraded documents.

Keywords: binarization, denoising, global thresholding, local thresholding, thresholding

Procedia PDF Downloads 302
4496 OCR/ICR Text Recognition Using ABBYY FineReader as an Example Text

Authors: A. R. Bagirzade, A. Sh. Najafova, S. M. Yessirkepova, E. S. Albert

Abstract:

This article describes a text recognition method based on Optical Character Recognition (OCR). The features of the OCR method were examined using the ABBYY FineReader program. It describes automatic text recognition in images. OCR is necessary because optical input devices can only transmit raster graphics as a result. Text recognition describes the task of recognizing letters shown as such, to identify and assign them an assigned numerical value in accordance with the usual text encoding (ASCII, Unicode). The peculiarity of this study conducted by the authors using the example of the ABBYY FineReader, was confirmed and shown in practice, the improvement of digital text recognition platforms developed by Electronic Publication.

Keywords: ABBYY FineReader system, algorithm symbol recognition, OCR/ICR techniques, recognition technologies

Procedia PDF Downloads 123
4495 Programmed Speech to Text Summarization Using Graph-Based Algorithm

Authors: Hamsini Pulugurtha, P. V. S. L. Jagadamba

Abstract:

Programmed Speech to Text and Text Summarization Using Graph-based Algorithms can be utilized in gatherings to get the short depiction of the gathering for future reference. This gives signature check utilizing Siamese neural organization to confirm the personality of the client and convert the client gave sound record which is in English into English text utilizing the discourse acknowledgment bundle given in python. At times just the outline of the gathering is required, the answer for this text rundown. Thus, the record is then summed up utilizing the regular language preparing approaches, for example, solo extractive text outline calculations

Keywords: Siamese neural network, English speech, English text, natural language processing, unsupervised extractive text summarization

Procedia PDF Downloads 168
4494 Reducing Accidents Using Text Stops

Authors: Benish Chaudhry

Abstract:

Most of the accidents these days are occurring because of the ‘text-and-drive’ concept. If we look at the structure of cities in UAE, there are great distances, because of which it is impossible to drive without using or merely checking the cellphone. Moreover, if we look at the road structure, it is almost impossible to stop at a point and text. With the introduction of TEXT STOPs, drivers will be able to stop different stops for a maximum of 1 and a half-minute in order to reply or write a message. They can be introduced at a distance of 10 minutes of driving on the average speed of the road, so the drivers can look forward to a stop and can reply to a text when needed. A user survey indicates that drivers are willing to NOT text-and-drive if they have such a facility available.

Keywords: transport, accidents, urban planning, road planning

Procedia PDF Downloads 337
4493 Structure Analysis of Text-Image Connection in Jalayrid Period Illustrated Manuscripts

Authors: Mahsa Khani Oushani

Abstract:

Text and image are two important elements in the field of Iranian art, the text component and the image component have always been manifested together. The image narrates the text and the text is the factor in the formation of the image and they are closely related to each other. The connection between text and image is an interactive and two-way connection in the tradition of Iranian manuscript arrangement. The interaction between the narrative description and the image scene is the result of a direct and close connection between the text and the image, which in addition to the decorative aspect, also has a descriptive aspect. In this article the connection between the text element and the image element and its adaptation to the theory of Roland Barthes, the structuralism theorist, in this regard will be discussed. This study tends to investigate the question of how the connection between text and image in illustrated manuscripts of the Jalayrid period is defined according to Barthes’ theory. And what kind of proportion has the artist created in the composition between text and image. Based on the results of reviewing the data of this study, it can be inferred that in the Jalayrid period, the image has a reference connection and although it is of major importance on the page, it also maintains a close connection with the text and is placed in a special proportion. It is not necessarily balanced and symmetrical and sometimes uses imbalance for composition. This research has been done by descriptive-analytical method, which has been done by library collection method.

Keywords: structure, text, image, Jalayrid, painter

Procedia PDF Downloads 179
4492 Text Localization in Fixed-Layout Documents Using Convolutional Networks in a Coarse-to-Fine Manner

Authors: Beier Zhu, Rui Zhang, Qi Song

Abstract:

Text contained within fixed-layout documents can be of great semantic value and so requires a high localization accuracy, such as ID cards, invoices, cheques, and passports. Recently, algorithms based on deep convolutional networks achieve high performance on text detection tasks. However, for text localization in fixed-layout documents, such algorithms detect word bounding boxes individually, which ignores the layout information. This paper presents a novel architecture built on convolutional neural networks (CNNs). A global text localization network and a regional bounding-box regression network are introduced to tackle the problem in a coarse-to-fine manner. The text localization network simultaneously locates word bounding points, which takes the layout information into account. The bounding-box regression network inputs the features pooled from arbitrarily sized RoIs and refine the localizations. These two networks share their convolutional features and are trained jointly. A typical type of fixed-layout documents: ID cards, is selected to evaluate the effectiveness of the proposed system. These networks are trained on data cropped from nature scene images, and synthetic data produced by a synthetic text generation engine. Experiments show that our approach locates high accuracy word bounding boxes and achieves state-of-the-art performance.

Keywords: bounding box regression, convolutional networks, fixed-layout documents, text localization

Procedia PDF Downloads 154
4491 Optimal Classifying and Extracting Fuzzy Relationship from Query Using Text Mining Techniques

Authors: Faisal Alshuwaier, Ali Areshey

Abstract:

Text mining techniques are generally applied for classifying the text, finding fuzzy relations and structures in data sets. This research provides plenty text mining capabilities. One common application is text classification and event extraction, which encompass deducing specific knowledge concerning incidents referred to in texts. The main contribution of this paper is the clarification of a concept graph generation mechanism, which is based on a text classification and optimal fuzzy relationship extraction. Furthermore, the work presented in this paper explains the application of fuzzy relationship extraction and branch and bound method to simplify the texts.

Keywords: extraction, max-prod, fuzzy relations, text mining, memberships, classification, memberships, classification

Procedia PDF Downloads 539
4490 Improving Subjective Bias Detection Using Bidirectional Encoder Representations from Transformers and Bidirectional Long Short-Term Memory

Authors: Ebipatei Victoria Tunyan, T. A. Cao, Cheol Young Ock

Abstract:

Detecting subjectively biased statements is a vital task. This is because this kind of bias, when present in the text or other forms of information dissemination media such as news, social media, scientific texts, and encyclopedias, can weaken trust in the information and stir conflicts amongst consumers. Subjective bias detection is also critical for many Natural Language Processing (NLP) tasks like sentiment analysis, opinion identification, and bias neutralization. Having a system that can adequately detect subjectivity in text will boost research in the above-mentioned areas significantly. It can also come in handy for platforms like Wikipedia, where the use of neutral language is of importance. The goal of this work is to identify the subjectively biased language in text on a sentence level. With machine learning, we can solve complex AI problems, making it a good fit for the problem of subjective bias detection. A key step in this approach is to train a classifier based on BERT (Bidirectional Encoder Representations from Transformers) as upstream model. BERT by itself can be used as a classifier; however, in this study, we use BERT as data preprocessor as well as an embedding generator for a Bi-LSTM (Bidirectional Long Short-Term Memory) network incorporated with attention mechanism. This approach produces a deeper and better classifier. We evaluate the effectiveness of our model using the Wiki Neutrality Corpus (WNC), which was compiled from Wikipedia edits that removed various biased instances from sentences as a benchmark dataset, with which we also compare our model to existing approaches. Experimental analysis indicates an improved performance, as our model achieved state-of-the-art accuracy in detecting subjective bias. This study focuses on the English language, but the model can be fine-tuned to accommodate other languages.

Keywords: subjective bias detection, machine learning, BERT–BiLSTM–Attention, text classification, natural language processing

Procedia PDF Downloads 89
4489 Audio-Visual Co-Data Processing Pipeline

Authors: Rita Chattopadhyay, Vivek Anand Thoutam

Abstract:

Speech is the most acceptable means of communication where we can quickly exchange our feelings and thoughts. Quite often, people can communicate orally but cannot interact or work with computers or devices. It’s easy and quick to give speech commands than typing commands to computers. In the same way, it’s easy listening to audio played from a device than extract output from computers or devices. Especially with Robotics being an emerging market with applications in warehouses, the hospitality industry, consumer electronics, assistive technology, etc., speech-based human-machine interaction is emerging as a lucrative feature for robot manufacturers. Considering this factor, the objective of this paper is to design the “Audio-Visual Co-Data Processing Pipeline.” This pipeline is an integrated version of Automatic speech recognition, a Natural language model for text understanding, object detection, and text-to-speech modules. There are many Deep Learning models for each type of the modules mentioned above, but OpenVINO Model Zoo models are used because the OpenVINO toolkit covers both computer vision and non-computer vision workloads across Intel hardware and maximizes performance, and accelerates application development. A speech command is given as input that has information about target objects to be detected and start and end times to extract the required interval from the video. Speech is converted to text using the Automatic speech recognition QuartzNet model. The summary is extracted from text using a natural language model Generative Pre-Trained Transformer-3 (GPT-3). Based on the summary, essential frames from the video are extracted, and the You Only Look Once (YOLO) object detection model detects You Only Look Once (YOLO) objects on these extracted frames. Frame numbers that have target objects (specified objects in the speech command) are saved as text. Finally, this text (frame numbers) is converted to speech using text to speech model and will be played from the device. This project is developed for 80 You Only Look Once (YOLO) labels, and the user can extract frames based on only one or two target labels. This pipeline can be extended for more than two target labels easily by making appropriate changes in the object detection module. This project is developed for four different speech command formats by including sample examples in the prompt used by Generative Pre-Trained Transformer-3 (GPT-3) model. Based on user preference, one can come up with a new speech command format by including some examples of the respective format in the prompt used by the Generative Pre-Trained Transformer-3 (GPT-3) model. This pipeline can be used in many projects like human-machine interface, human-robot interaction, and surveillance through speech commands. All object detection projects can be upgraded using this pipeline so that one can give speech commands and output is played from the device.

Keywords: OpenVINO, automatic speech recognition, natural language processing, object detection, text to speech

Procedia PDF Downloads 38
4488 Emotional Analysis for Text Search Queries on Internet

Authors: Gemma García López

Abstract:

The goal of this study is to analyze if search queries carried out in search engines such as Google, can offer emotional information about the user that performs them. Knowing the emotional state in which the Internet user is located can be a key to achieve the maximum personalization of content and the detection of worrying behaviors. For this, two studies were carried out using tools with advanced natural language processing techniques. The first study determines if a query can be classified as positive, negative or neutral, while the second study extracts emotional content from words and applies the categorical and dimensional models for the representation of emotions. In addition, we use search queries in Spanish and English to establish similarities and differences between two languages. The results revealed that text search queries performed by users on the Internet can be classified emotionally. This allows us to better understand the emotional state of the user at the time of the search, which could involve adapting the technology and personalizing the responses to different emotional states.

Keywords: emotion classification, text search queries, emotional analysis, sentiment analysis in text, natural language processing

Procedia PDF Downloads 104
4487 Mask-Prompt-Rerank: An Unsupervised Method for Text Sentiment Transfer

Authors: Yufen Qin

Abstract:

Text sentiment transfer is an important branch of text style transfer. The goal is to generate text with another sentiment attribute based on a text with a specific sentiment attribute while maintaining the content and semantic information unrelated to sentiment unchanged in the process. There are currently two main challenges in this field: no parallel corpus and text attribute entanglement. In response to the above problems, this paper proposed a novel solution: Mask-Prompt-Rerank. Use the method of masking the sentiment words and then using prompt regeneration to transfer the sentence sentiment. Experiments on two sentiment benchmark datasets and one formality transfer benchmark dataset show that this approach makes the performance of small pre-trained language models comparable to that of the most advanced large models, while consuming two orders of magnitude less computing and memory.

Keywords: language model, natural language processing, prompt, text sentiment transfer

Procedia PDF Downloads 38
4486 Exploratory Analysis of A Review of Nonexistence Polarity in Native Speech

Authors: Deawan Rakin Ahamed Remal, Sinthia Chowdhury, Sharun Akter Khushbu, Sheak Rashed Haider Noori

Abstract:

Native Speech to text synthesis has its own leverage for the purpose of mankind. The extensive nature of art to speaking different accents is common but the purpose of communication between two different accent types of people is quite difficult. This problem will be motivated by the extraction of the wrong perception of language meaning. Thus, many existing automatic speech recognition has been placed to detect text. Overall study of this paper mentions a review of NSTTR (Native Speech Text to Text Recognition) synthesis compared with Text to Text recognition. Review has exposed many text to text recognition systems that are at a very early stage to comply with the system by native speech recognition. Many discussions started about the progression of chatbots, linguistic theory another is rule based approach. In the Recent years Deep learning is an overwhelming chapter for text to text learning to detect language nature. To the best of our knowledge, In the sub continent a huge number of people speak in Bangla language but they have different accents in different regions therefore study has been elaborate contradictory discussion achievement of existing works and findings of future needs in Bangla language acoustic accent.

Keywords: TTR, NSTTR, text to text recognition, deep learning, natural language processing

Procedia PDF Downloads 89
4485 DCDNet: Lightweight Document Corner Detection Network Based on Attention Mechanism

Authors: Kun Xu, Yuan Xu, Jia Qiao

Abstract:

The document detection plays an important role in optical character recognition and text analysis. Because the traditional detection methods have weak generalization ability, and deep neural network has complex structure and large number of parameters, which cannot be well applied in mobile devices, this paper proposes a lightweight Document Corner Detection Network (DCDNet). DCDNet is a two-stage architecture. The first stage with Encoder-Decoder structure adopts depthwise separable convolution to greatly reduce the network parameters. After introducing the Feature Attention Union (FAU) module, the second stage enhances the feature information of spatial and channel dim and adaptively adjusts the size of receptive field to enhance the feature expression ability of the model. Aiming at solving the problem of the large difference in the number of pixel distribution between corner and non-corner, Weighted Binary Cross Entropy Loss (WBCE Loss) is proposed to define corner detection problem as a classification problem to make the training process more efficient. In order to make up for the lack of Dataset of document corner detection, a Dataset containing 6620 images named Document Corner Detection Dataset (DCDD) is made. Experimental results show that the proposed method can obtain fast, stable and accurate detection results on DCDD.

Keywords: document detection, corner detection, attention mechanism, lightweight

Procedia PDF Downloads 317
4484 Adaptation of Hough Transform Algorithm for Text Document Skew Angle Detection

Authors: Kayode A. Olaniyi, Olabanji F. Omotoye, Adeola A. Ogunleye

Abstract:

The skew detection and correction form an important part of digital document analysis. This is because uncompensated skew can deteriorate document features and can complicate further document image processing steps. Efficient text document analysis and digitization can rarely be achieved when a document is skewed even at a small angle. Once the documents have been digitized through the scanning system and binarization also achieved, document skew correction is required before further image analysis. Research efforts have been put in this area with algorithms developed to eliminate document skew. Skew angle correction algorithms can be compared based on performance criteria. Most important performance criteria are accuracy of skew angle detection, range of skew angle for detection, speed of processing the image, computational complexity and consequently memory space used. The standard Hough Transform has successfully been implemented for text documentation skew angle estimation application. However, the standard Hough Transform algorithm level of accuracy depends largely on how much fine the step size for the angle used. This consequently consumes more time and memory space for increase accuracy and, especially where number of pixels is considerable large. Whenever the Hough transform is used, there is always a tradeoff between accuracy and speed. So a more efficient solution is needed that optimizes space as well as time. In this paper, an improved Hough transform (HT) technique that optimizes space as well as time to robustly detect document skew is presented. The modified algorithm of Hough Transform presents solution to the contradiction between the memory space, running time and accuracy. Our algorithm starts with the first step of angle estimation accurate up to zero decimal place using the standard Hough Transform algorithm achieving minimal running time and space but lacks relative accuracy. Then to increase accuracy, suppose estimated angle found using the basic Hough algorithm is x degree, we then run again basic algorithm from range between ±x degrees with accuracy of one decimal place. Same process is iterated till level of desired accuracy is achieved. The procedure of our skew estimation and correction algorithm of text images is implemented using MATLAB. The memory space estimation and process time are also tabulated with skew angle assumption of within 00 and 450. The simulation results which is demonstrated in Matlab show the high performance of our algorithms with less computational time and memory space used in detecting document skew for a variety of documents with different levels of complexity.

Keywords: hough-transform, skew-detection, skew-angle, skew-correction, text-document

Procedia PDF Downloads 118