Vqa and Visual Reasoning: An Overview of Approaches, Datasets, and Future Direction

Kód výsledku v IS VaVaI

<a href="https://www.isvavai.cz/riv?ss=detail&h=RIV%2F00216208%3A11320%2F23%3ASJSEXNPY" target="_blank" >RIV/00216208:11320/23:SJSEXNPY - isvavai.cz</a>

Výsledek na webu

<a href="https://papers.ssrn.com/sol3/papers.cfm?abstract_id=4455698" target="_blank" >https://papers.ssrn.com/sol3/papers.cfm?abstract_id=4455698</a>

DOI - Digital Object Identifier

<a href="http://dx.doi.org/10.2139/ssrn.4455698" target="_blank" >10.2139/ssrn.4455698</a>

Jazyk výsledku

angličtina

Název v původním jazyce

Vqa and Visual Reasoning: An Overview of Approaches, Datasets, and Future Direction

Popis výsledku v původním jazyce

"Visual Question Answering (VQA) is a challenging and fascinating area of research in computer vision, deep learning, and natural language processing that has recently piqued the interest of AI researchers from a variety of fields. The inputs for this task are an image and a written question about the image. The answer can be generated intelligently or selected from a set of options related to the image. We have seen a rapid development of VQA and visual reasoning in tasks based on deep learning and massive amounts of annotated data over the last five years. In this article, we will present a comprehensive literature review of the current state of the art in VQA and visual reasoning from four perspectives: problem definition and challenges, approaches, existing datasets, and evaluation matrices. We discussed the datasets' limitations and thoroughly examined the current models. Finally, we investigate potential future research directions in this area in order to generate new ideas and creative approaches to solving current problems and developing new applications."

Název v anglickém jazyce

Vqa and Visual Reasoning: An Overview of Approaches, Datasets, and Future Direction

Popis výsledku anglicky

"Visual Question Answering (VQA) is a challenging and fascinating area of research in computer vision, deep learning, and natural language processing that has recently piqued the interest of AI researchers from a variety of fields. The inputs for this task are an image and a written question about the image. The answer can be generated intelligently or selected from a set of options related to the image. We have seen a rapid development of VQA and visual reasoning in tasks based on deep learning and massive amounts of annotated data over the last five years. In this article, we will present a comprehensive literature review of the current state of the art in VQA and visual reasoning from four perspectives: problem definition and challenges, approaches, existing datasets, and evaluation matrices. We discussed the datasets' limitations and thoroughly examined the current models. Finally, we investigate potential future research directions in this area in order to generate new ideas and creative approaches to solving current problems and developing new applications."

Druh

O - Ostatní výsledky

CEP obor

—

OECD FORD obor

10201 - Computer sciences, information science, bioinformathics (hardware development to be 2.2, social aspect to be 5.8)

Projekt

—

Návaznosti

—

Rok uplatnění

2023

Kód důvěrnosti údajů

S - Úplné a pravdivé údaje o projektu nepodléhají ochraně podle zvláštních právních předpisů