Image and Video Captioning with Augmented Neural Architectures

Rakshith Shetty; Hamed Rezazadegan Tavakoli; Jorma Laaksonen

doi:10.1109/MMUL.2018.112135923

Image and Video Captioning with Augmented Neural Architectures

Rakshith Shetty, Hamed Rezazadegan Tavakoli, Jorma Laaksonen

Tutkimustuotos: Lehtiartikkeli › Article › Scientific › vertaisarvioitu

Abstrakti

Neural-network-based image and video captioning can be substantially improved by utilizing architectures that make use of special features from the scene context, objects, and locations. A novel discriminatively trained evaluator network for choosing the best caption among those generated by an ensemble of caption generator networks further improves accuracy.

Alkuperäiskieli	Englanti
Sivut	34-46
Sivumäärä	13
Julkaisu	IEEE Multimedia
Vuosikerta	25
Numero	2
DOI - pysyväislinkit	https://doi.org/10.1109/MMUL.2018.112135923
Tila	Julkaistu - 2018
OKM-julkaisutyyppi	A1 Julkaistu artikkeli, soviteltu

Pääsy asiakirjaan

10.1109/MMUL.2018.112135923

Sormenjälki Sukella tutkimusaiheisiin 'Image and Video Captioning with Augmented Neural Architectures'. Ne muodostavat yhdessä ainutlaatuisen sormenjäljen.

Laitteet

Science-IT

Mikko Hakala (Manager)

Perustieteiden korkeakoulu

Laitteistot/tilat: Facility

Siteeraa tätä

Shetty, R., Rezazadegan Tavakoli, H., & Laaksonen, J. (2018). Image and Video Captioning with Augmented Neural Architectures. IEEE Multimedia, 25(2), 34-46. https://doi.org/10.1109/MMUL.2018.112135923

@article{a36e040a6ac44cb2b660f6c231c40a19,

title = "Image and Video Captioning with Augmented Neural Architectures",

abstract = "Neural-network-based image and video captioning can be substantially improved by utilizing architectures that make use of special features from the scene context, objects, and locations. A novel discriminatively trained evaluator network for choosing the best caption among those generated by an ensemble of caption generator networks further improves accuracy.",

keywords = "computer vision, applications and expert knowledge-intensive systems, artificial intelligence, computing, deep learning, image captioning, recurrent networks ",

author = "Rakshith Shetty and {Rezazadegan Tavakoli}, Hamed and Jorma Laaksonen",

year = "2018",

doi = "10.1109/MMUL.2018.112135923",

language = "English",

volume = "25",

pages = "34--46",

journal = "IEEE Multimedia",

issn = "1070-986X",

publisher = "IEEE Computer Society",

number = "2",

}

TY - JOUR

T1 - Image and Video Captioning with Augmented Neural Architectures

AU - Shetty, Rakshith

AU - Rezazadegan Tavakoli, Hamed

AU - Laaksonen, Jorma

PY - 2018

Y1 - 2018

N2 - Neural-network-based image and video captioning can be substantially improved by utilizing architectures that make use of special features from the scene context, objects, and locations. A novel discriminatively trained evaluator network for choosing the best caption among those generated by an ensemble of caption generator networks further improves accuracy.

AB - Neural-network-based image and video captioning can be substantially improved by utilizing architectures that make use of special features from the scene context, objects, and locations. A novel discriminatively trained evaluator network for choosing the best caption among those generated by an ensemble of caption generator networks further improves accuracy.

KW - computer vision

KW - applications and expert knowledge-intensive systems

KW - artificial intelligence

KW - computing

KW - deep learning

KW - image captioning

KW - recurrent networks

U2 - 10.1109/MMUL.2018.112135923

DO - 10.1109/MMUL.2018.112135923

M3 - Article

VL - 25

SP - 34

EP - 46

JO - IEEE Multimedia

JF - IEEE Multimedia

SN - 1070-986X

IS - 2

ER -