Loughborough University
Browse
On the Limitations of Visual-Semantic Embedding Networks for Image-to-Text Information Retrieval _ Enhanced Reader.pdf (9.52 MB)

On the limitations of visual-semantic embedding networks for image-to-text information retrieval

Download (9.52 MB)
journal contribution
posted on 2021-08-02, 10:35 authored by Yan Gong, Georgina CosmaGeorgina Cosma, Hui FangHui Fang
Visual-semantic embedding (VSE) networks create joint image–text representations to map images and texts in a shared embedding space to enable various information retrieval-related tasks, such as image–text retrieval, image captioning, and visual question answering. The most recent state-of-the-art VSE-based networks are: VSE++, SCAN, VSRN, and UNITER. This study evaluates the performance of those VSE networks for the task of image-to-text retrieval and identifies and analyses their strengths and limitations to guide future research on the topic. The experimental results on Flickr30K revealed that the pre-trained network, UNITER, achieved 61.5% on average Recall@5 for the task of retrieving all relevant descriptions. The traditional networks, VSRN, SCAN, and VSE++, achieved 50.3%, 47.1%, and 29.4% on average Recall@5, respectively, for the same task. An additional analysis was performed on image–text pairs from the top 25 worst-performing classes using a subset of the Flickr30K-based dataset to identify the limitations of the performance of the best-performing models, VSRN and UNITER. These limitations are discussed from the perspective of image scenes, image objects, image semantics, and basic functions of neural networks. This paper discusses the strengths and limitations of VSE networks to guide further research into the topic of using VSE networks for cross-modal information retrieval tasks.

History

School

  • Science

Department

  • Computer Science

Published in

Journal of Imaging

Volume

7

Issue

8

Publisher

MDPI AG

Version

  • VoR (Version of Record)

Rights holder

© The Authors

Publisher statement

This is an Open Access Article. It is published my MDPI under the Creative Commons Attribution 4.0 International Licence (CC BY 4.0). Full details of this licence are available at: https://creativecommons.org/licenses/by/4.0/

Acceptance date

2021-07-23

Publication date

2021-07-26

Copyright date

2021

eISSN

2313-433X

Language

  • en

Depositor

Dr Georgina Cosma. Deposit date: 27 July 2021

Article number

125

Usage metrics

    Loughborough Publications

    Categories

    No categories selected

    Licence

    Exports

    RefWorks
    BibTeX
    Ref. manager
    Endnote
    DataCite
    NLM
    DC