Exploring the Relationship between Dataset Size and Image Captioning Model Performance

Kód výsledku v IS VaVaI

<a href="https://www.isvavai.cz/riv?ss=detail&h=RIV%2F49777513%3A23520%2F23%3A43969948" target="_blank" >RIV/49777513:23520/23:43969948 - isvavai.cz</a>

Výsledek na webu

<a href="http://svk.fav.zcu.cz/download/proceedings_svk_2023.pdf" target="_blank" >http://svk.fav.zcu.cz/download/proceedings_svk_2023.pdf</a>

DOI - Digital Object Identifier

—

Jazyk výsledku

angličtina

Název v původním jazyce

Exploring the Relationship between Dataset Size and Image Captioning Model Performance

Popis výsledku v původním jazyce

Image captioning is a deep learning task, which goal is to automatically generate textual description of an input image. It is a complex task that requires identifying and interpreting visual information and generating grammatically correct and fluent sentences. Because different individuals may consider various aspects of an image important, there isn’t any single correct caption. This means that there is no ideal evaluation metric for measuring caption quality, as different metrics may better evaluate different attributes of the caption. Image captioning models, just like other deep learning models, need a large amount of training data and require a long time to train. In this work, we investigate the impact of using a smaller amount of training data on the performance of the standard image captioning model Oscar.

Název v anglickém jazyce

Exploring the Relationship between Dataset Size and Image Captioning Model Performance

Popis výsledku anglicky

Image captioning is a deep learning task, which goal is to automatically generate textual description of an input image. It is a complex task that requires identifying and interpreting visual information and generating grammatically correct and fluent sentences. Because different individuals may consider various aspects of an image important, there isn’t any single correct caption. This means that there is no ideal evaluation metric for measuring caption quality, as different metrics may better evaluate different attributes of the caption. Image captioning models, just like other deep learning models, need a large amount of training data and require a long time to train. In this work, we investigate the impact of using a smaller amount of training data on the performance of the standard image captioning model Oscar.

Druh

O - Ostatní výsledky

CEP obor

—

OECD FORD obor

20205 - Automation and control systems

Projekt

—

Návaznosti

S - Specificky vyzkum na vysokych skolach

Rok uplatnění

2023

Kód důvěrnosti údajů

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