Is this you? Claim this profile to correct it, add a bio and choose the work people see first.
Claim this profileAcademic lineage
Possible advisorsa guess from early papers, not confirmed
- Apostolos AntonacopoulosSuggested from co-authorship
Is this you? Claim this profile to confirm or dismiss it.
Works215 from public data
- DocVQA: A Dataset for VQA on Document Images1,800
Although the existing models perform reasonably well on certain types of questions, there is large performance gap compared to human performance (94.36% accuracy).
- 1,608
- 1,492
- InfographicVQA564
The results suggest that VQA on infographics—images that are designed to communicate information quickly and clearly to human brain—is ideal for benchmarking machine understanding of complex document images.
- Scene Text Visual Question Answering545
A new dataset, ST-VQA, is presented that aims to highlight the importance of exploiting high-level semantic information present in images as textual cues in the Visual Question Answering process and proposes a new evaluation metric for these tasks to account both for reasoning errors as well as shortcomings of the text recognition module.
- 455
- ICDAR2019 Competition on Scanned Receipt OCR and Information Extraction452
The ICDAR 2019 Challenge on "Scanned receipts OCR and key information extraction" (SROIE) covers important aspects related to the automated analysis of scanned receipts, and is considered to evolve into a useful resource for the community, drawing further attention and promoting research and development efforts in this field.
- Exploring Hate Speech Detection in Multimodal Publications322
It is found that, even though images are useful for the hate speech detection task, current multimodal models cannot outperform models analyzing only text.
- ICDAR2019 Robust Reading Challenge on Multi-lingual Scene Text Detection and Recognition — RRC-MLT-2019301
The dataset, the tasks and the findings of the presented RRC-MLT-2019 challenge are presented, which has 4 tasks covering various aspects of multi-lingual scene text.
- ICDAR2019 Robust Reading Challenge on Arbitrary-Shaped Text - RRC-ArT272
This paper reports the ICDAR2019 Robust Reading Challenge on Arbitrary-Shaped Text - RRC-ArT that consists of three major challenges: i) scene text detection, ii) sceneText recognition, and iii) scenetext spotting.
- ICDAR 2019 Robust Reading Challenge on Reading Chinese Text on Signboard209
A large-scale dataset of 25,000 annotated signboard images, in which all the text lines and characters are annotated with locations and transcriptions, were released and a multi ground truth (multi-GT) evaluation method was proposed to make evaluation fairer.
- ICDAR 2019 Competition on Large-Scale Street View Text with Partial Labeling - RRC-LSVT207
This competition aims to explore the abilities of state-of-the-art methods to detect and recognize text instances from large-scale street view images, closing gaps between research benchmarks and real applications.
- Good News, Everyone! Context Driven Entity-Aware Captioning for News Images161
This work proposes a novel captioning method able to leverage contextual information provided by the text of news articles associated with an image, able to selectively draw information from the article guided by visual cues, and to dynamically extend the output dictionary to out-of-vocabulary named entities that appear in the context source.
- Self-Supervised Learning of Visual Features through Embedding Images into Text Topic Spaces130
It is shown that discriminative visual features can be learnt efficiently by training a CNN to predict the semantic context in which a particular image is more probable to appear as an illustration, and the hidden semantic structures discovered in the text corpus are leverage with a well-known topic modeling technique.
- TextProposals: A text-specific selective search algorithm for word spotting in the wild114
A novel object proposals method that is specifically designed for text, based on a similarity based region grouping algorithm that generates a hierarchy of word hypotheses, that shows competitive performance in end-to-end word spotting and outperforms previously published results.
- 111
- 108
- Let there be a clock on the beach: Reducing Object Hallucination in Image Captioning104
By extensive analysis, it is shown that the proposed methods can significantly diminish the authors' models’ object bias on hallucination metrics and experimentally demonstrate that the methods decrease the dependency on the visual features.
- Generation of synthetic documents for performance evaluation of symbol recognition & spotting systems98
A new approach to the generation of synthetic graphics documents containing non-isolated symbols in a real context based on the definition of a set of constraints that permits to generate documents with different features that are reflected in variation of localization results.
- ICDAR 2019 Competition on Scene Text Visual Question Answering91
This paper presents final results of ICDAR 2019 Scene Text Visual Question Answering competition (ST-VQA), which introduces a new dataset comprising 23,038 images annotated with 31,791 question / answer pairs where the answer is always grounded on text instances present in the image.
- Improving patch-based scene text script identification with ensembles of conjoined networks90
A patch-based classification method for script identificattion in the wild based on the use of ensembles of conjoined networks and a new public benchmark dataset for the evaluation of multi-lingual scene text end-to-end reading systems.
- 73
- 70
- 64
- 62
- 61
- 59
- Document Collection Visual Question Answering56
This work introduces Document Collection Visual Question Answering (DocCVQA) a new dataset and related task, where questions are posed over a whole collection of document images and the goal is not only to provide the answer to the given question, but also to retrieve the set of documents that contain the information needed to infer the answer.
- Single Shot Scene Text Retrieval55
This paper addresses the problem of scene text retrieval: given a text query, the system must return all images containing the queried text and proposes a single shot CNN architecture that predicts at the same time bounding boxes and a compact text representation of the words in them.
- Sparse radial sampling LBP for writer identification53
The proposed method is the only one that is based on dense extraction of a single local feature descriptor that makes it fast and applicable at the earliest stages in a DIA pipeline without the need for segmentation, binarization, or extraction of multiple features.
- The achromatic locus: Effect of navigation direction in color space52
Observers are able to make more reliable achromatic settings when adjusting the test patch along the directions defined by the four unique hues as opposed to navigating along the main axes in the commonly used CIE u*v* chromaticity plane.
- A Fine-Grained Approach to Scene Text Script Identification49
A novel method for script identification in natural images that combines convolutional features and the Naive-Bayes Nearest Neighbor classifier, which efficiently exploits the discriminative power of small stroke-parts, in a fine-grained classification framework is detailed.
- Knowledge-driven understanding of images in comic books49
This study proposes a knowledge-driven system that can interact with bottom-up and top-down information to progressively understand the content of a document and model the comic book and the image processing domains knowledge for information consistency analysis.
- 49
- OCR-IDL: OCR Annotations for Industry Document Library Dataset48
This work makes public the OCR annotations for IDL documents using commercial OCR engine given their superior performance over open source OCR models.
- Colour text segmentation in web images based on human perception47
The proposed approach enables the segmentation of text in complex situations such as in the presence of varying colour and texture (characters and background), and approximate the human colour perception characteristics for the identification of character regions.
- A framework for the assessment of text extraction algorithms on complex colour images45
A complete framework for the evaluation of text extraction methods at multiple levels is described, a detailed ground-truth specification is provided and a case study is presented on how this framework can be used in a real-life situation.
- A fast hierarchical method for multi-script and arbitrary oriented scene text extraction42
This paper addresses the problem of text segmentation in natural scenes from a hierarchical perspective introducing a feature space designed to produce text group hypotheses with high recall and a novel stopping rule combining a discriminative classifier and a probabilistic measure of group meaningfulness based on perceptual organization.
- Ground Truth for Layout Analysis Performance Evaluation41
The effectiveness of the ground truth discussed in this paper has been successfully shown in its use for two international page segmentation competitions (ICDAR2003 and ICDAR2005).
- 39
- Multi-Modal Reasoning Graph for Scene-Text Based Fine-Grained Image Classification and Retrieval38
This paper employs a Graph Convolutional Network to perform multi-modal reasoning and obtain relationship-enhanced features by learning a common semantic space between salient objects and text found in an image.
- Multimodal page classification in administrative document image streams38
This paper presents a page classification application in a banking workflow that represents administrative document images by merging visual and textual descriptions and reports results on a dataset of 70,000 pages.
- Text-DIAE: A Self-Supervised Degradation Invariant Autoencoder for Text Recognition and Document Enhancement37
This paper proposes a Text-Degradation Invariant Auto Encoder (Text-DIAE), a self-supervised model designed to tackle two tasks, text recognition (handwritten or scene-text) and document image enhancement, and demonstrates that the method surpasses the state-of-the-art in existing supervised and self- supervised settings in handwritten and scene text recognition and document images enhancement.
- 36
- Accessing textual information embedded in Internet images36
Research under way to devise tools to extracted text from images based on the way humans perceive color differences is outlined and results are presented.
- ICDAR 2021 Competition on Document Visual Question Answering34
Results of the ICDAR 2021 edition of the Document Visual Question Challenges are presented, including a newly introduced on Infographics VQA, based on a new dataset of more than 5,000 infographics images and 30,000 question-answer pairs.
- Fine-grained Image Classification and Retrieval by Combining Visual and Locally Pooled Textual Features34
This approach provides a stronger multimodal representation for this task and as the experiments demonstrate, it achieves state-of-the-art results on two different tasks, fine-grained classification and image retrieval.
- 34
- 34
- 31
- 30
- Automatic Text Localisation in Scanned Comic Books29
A novel approach for the automatic text localization in scanned comics book pages, an essential step towards a fully automatic comics book understanding, focuses on speech text as it is semantically important and represents the majority of the text present in comics.
- 29
- StacMR: Scene-Text Aware Cross-Modal Retrieval28
This paper proposes a new dataset that allows exploration of cross-modal retrieval where images contain scene-text instances and describes several approaches which leverage scene text, including a better scene- text aware cross- modal retrieval method which uses specialized representations for text from the captions and text fromThe visual scene.
- Self-Supervised Learning from Web Data for Multimodal Retrieval28
It is shown that the embeddings learned with web and social media data have competitive performances over supervised methods in the text-based image retrieval task, and clearly outperform the state of the art in the MIRFlickr dataset when training in the target data.
- A generic framework for median graph computation based on a recursive embedding approach28
This paper proposes a generic recursive procedure that permits to recover the graph corresponding to a point in the vector space, introducing only the amount of approximation inherent to the use of graph matching algorithms.
- Visual gamma correction for LCD displays28
This observer-based method of visual gamma correction is useful to estimate the OETF for LCD displays and its major advantage is that no particular functional relationship between digital inputs and luminance outputs has to be assumed.
- 27
- 27
- Rotation invariant hand-drawn symbol recognition based on a dynamic time warping model27
This paper introduces a robust approach for describing and recognizing hand-drawn symbols tolerant to these writer style differences based on the dynamic time warping (DTW) algorithm.
- Watching the News: Towards VideoQA Models that can Read26
This work proposes a novel VideoQA task that requires reading and understanding the text in the video and presents the "NewsVideoQA" dataset, a dataset that comprises more than 8, 600 QA pairs on 3, 000+ news videos obtained from diverse news channels from around the world.
- Real-time Lexicon-free Scene Text Retrieval26
The proposed model uses a single shot CNN architecture that predicts bounding boxes and builds a compact representation of spotted words, which can be modeled as a nearest neighbor search of the textual representation of a query over the outputs of the CNN collected from the totality of an image database.
- Learning to Learn from Web Data Through Deep Semantic Embeddings25
It is shown that the embeddings learnt with Web and Social Media data have competitive performances over supervised methods in the text based image retrieval task, and they clearly outperform state of the art in the MIRFlickr dataset when training in the target data.
- 24
- Multimodal grid features and cell pointers for scene text visual question answering23
A new model is presented for the task of scene text visual question answering, in which questions about a given image can only be answered by reading and understanding scene text that is present in it, based on an attention mechanism that attends to multi-modal features conditioned to the question.
- 23
- Scene Text Recognition: No Country for Old Men?23
This paper confirms that a conventional shape-based OCR engine would be able to produce competitive results in the end-to-end scene text recognition task when provided with a conveniently preprocessed image.
- Object proposals for text extraction in the wild22
A new Object Proposals algorithm that is specifically designed for text is proposed that is superior in its ability of producing good quality word proposals in an efficient way and is compared with other generic methods in the state of the art.
- The lifecycle of a digital historical document22
Tests indicate the superior performance of the whole semantics-driven approach both over manual transcription and over the semi-automated application of off-the-shelf OCR and the use of a conventional (text and layout) document format.
- An Anthropocentric Approach to Text Extraction from WWW Images22
The method described here is the first in the systematic approach to approximate the human colour perception characteristics for the identification of character regions and is decomposed by performing histogram analysis of Hue and Luminance and merging in the HLS colour space.
- Visual Script and Language Identification21
A script identification method based on hand-crafted texture features and an artificial neural network that allows the use of discriminative models on unknown classes and demonstrates near state-of-the-art performance on visual language identification of handwritten text.
- Is An Image Worth Five Sentences? A New Look into Semantics for Image-Text Matching20
This work proposes two metrics that evaluate the degree of semantic relevance of retrieved items, independently of their annotated binary relevance, and incorporates a novel strategy that uses an image captioning metric, CIDEr, to define a Semantic Adaptive Margin (SAM) to be optimized in a standard triplet loss.
- Learning from #Barcelona Instagram Data What Locals and Tourists Post About Its Neighbourhoods20
This work gathers Instagram data related to Barcelona consisting on images-captions pairs and uses the text as a supervisory signal to learn relations between images, words and neighborhoods, and shows that tourism is reflected in Social Media at a neighborhood level.
- FAST: Facilitated and Accurate Scene Text Proposals through FCN Guided Pruning20
This paper combines the Text Proposals algorithm with Fully Convolutional Networks to efficiently reduce the number of proposals while maintaining the same recall level and thus gaining a significant speed up.
- 20
- Interactive Trademark Image Retrieval by Fusing Semantic and Visual Content19
An efficient queried-by-example retrieval system which is able to retrieve trademark images by similarity from patent and trademark offices' digital libraries is proposed.
- A Complete Approach to the Conversion of Typewritten Historical Documents for Digital Archives19
A complete system that historians/archivists can use to digitize whole collections of documents relating to personal information, which shows a significant improvement not only in terms of OCR accuracy but also in overall time/cost involved in converting these documents for digital archives.
- Asking questions on handwritten document collections18
It is argued that the recognition-free approach is suitable for handwritten documents and historical collections where robust text recognition is often difficult and for human users, document image snippets containing answers act as a valid alternative to textual answers.
- Self-Supervised Visual Representations for Cross-Modal Retrieval18
The experiments demonstrate that the proposed method is not only capable of learning discriminative visual representations for solving vision tasks like classification, but that the learned representations are better for cross-modal retrieval when compared to supervised pre-training of the network on the ImageNet dataset.
- 18
- Colour‐opponent mechanisms are not affected by age‐related chromatic sensitivity changes18
A significant decrease in chromatic sensitivity due to ageing is found, in particular along the tritan line, which provides evidence in favour of the hypothesis that higher‐order colour mechanisms are equipped with flexible cone weightings, as opposed to fixed weights.
- Selective Style Transfer for Text17
Results on different text domains and cross-modal results demonstrate that this is feasible, and open different research lines, and two architectures for selective style transfer, which means transferring style to only desired image pixels are proposed.
- Dynamic Lexicon Generation for Natural Scene Images17
This paper proposes a method that generates contextualized lexicons for scene images using only visual information and demonstrates that the quality of the automatically obtained custom lexicons is superior to a generic frequency-based baseline.
- Modelling Task-Dependent Eye Guidance to Objects in Pictures17
A model of attentional eye guidance is introduced based on the rationale that the deployment of gaze is to be considered in the context of a general action-perception loop relying on two strictly intertwined processes.
- 16
- 16
- One-shot Compositional Data Generation for Low Resource Handwritten Text Recognition15
Contrary to traditional generation approaches, which require a huge amount of annotated images, the proposed method is able to generate human-like handwriting using only one sample of each symbol from the desired alphabet.
- 15
- 15
- Unique hue data for colour appearance models. Part II: Chromatic adaptation transform15
Using the large dataset of unique hue data as a benchmark, an optimised parameter is proposed for chromatic adaptation under mixed illumination conditions that produces more accurate results in unique hue reproduction.
- Two approaches for text segmentation in Web images15
Two novel methods for the segmentation of characters for subsequent extraction and recognition are presented and the novelty of both approaches is the combination of topological features of characters with an anthropocentric perspective of colour perception - in preference to RGB space analysis.
- Spotting Graphical Symbols in Camera-Acquired Documents in Real Time14
The system is based on the extraction and further matching of ORB compact local features computed over interest key-points and the FLANN indexing framework based on approximate nearest neighbor search allows to efficiently match local descriptors between the captured scene and the graphical models.
- 14
- 13
- Fuzzy Segmentation of Characters in Web Images Based on Human Colour Perception12
A new approach for the segmentation of characters in images on Web pages that attempts to emulate the ability of humans to differentiate between colours using a colour distance defined in a perceptually uniform colour space.
- 11
- 11
- 10
- DSText V2: A Comprehensive Video Text Spotting Dataset for Dense and Small Text10
A video text reading benchmark, DSText is established, which focuses on dense and small text reading challenges in the video with various scenarios, and it is hoped the benchmark will promise video text research in the community.
- Limitations of visual gamma corrections in LCD displays10
The results show that there are significant differences between the gamma transfer functions measured and produced by observers at different viewing angles, and suggest appropriate modifications to the Xiao et al. paradigm to counterbalance these artefacts.
- 9
- 9
- 9
- 9
- 9
- 8
- 7
- 7
- 7
- 7
- 7
- 7
- 7
- 7
- 6
- 6
- Text Recognition - Real World Data and Where to Find Them6
It is shown that training with the extracted PGT consistently improves the accuracy of a state of the art recognition model, by 3.7% on average, across different benchmark datasets (image domains) and 24.5% on one of the weakly annotated datasets.
- 6
- 5
- 5
- 5
- 4
- 4
- 4
- 4
- 4
- 4
- 3
- 3
- 3
- 3
- 3
- 3
- 3
- 3
- 3
- 3
- 3
- 3
- 3
- 3
- 3
- 3
- 3
- 2
- 2
- 2
- 2
- 2
- 2
- 2
- 2
- 2
- 2
- 2
- 2
- 1
- 1
- 1
- 1
- 1
- 1
- 1
- 1
- 1
- 1
- 1
- 1
- 1
- 1
- 1
- 1
- Rethinking Expert Training for Model Merging with Prompt Learning–
Dual-Tuned Experts (DTEs), a two-stage training strategy that first learns prompts and then fine-tunes the vision encoder, is introduced, a two-stage training strategy that reduces the magnitude of task-specific parameter updates and produces experts with higher merge compatibility.
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- A Benchmark for Symbolic Reasoning from Pixel Sequences: Grid-Level Visual Completion and Correction–
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
Publication data from OpenAlex, with missing venues and authors filled in from Crossref; citation counts are the higher of OpenAlex and Semantic Scholar, last synced 2026-10-16. One-sentence summaries under some papers are written by Semantic Scholar’s model. Citation counts may be lower than on Google Scholar, which indexes more sources.
Report an error
Wrong papers, two people merged into one, or a profile that should not be here? Tell us and we will fix or hide it.