Cross Encoder-Decoder Transformer with Global-Local Visual Extractor for Medical Image Captioning
Transformer-based approaches have shown good results in image captioning tasks. However, current approaches have a limitation in generating text from global features of an entire image. Therefore, we propose novel methods for generating better image captioning as follows: (1) The Global-Local Visual...
Main Authors: | , , , , |
---|---|
Format: | Article |
Language: | English |
Published: |
MDPI AG
2022-02-01
|
Series: | Sensors |
Subjects: | |
Online Access: | https://www.mdpi.com/1424-8220/22/4/1429 |
_version_ | 1797476726816636928 |
---|---|
author | Hojun Lee Hyunjun Cho Jieun Park Jinyeong Chae Jihie Kim |
author_facet | Hojun Lee Hyunjun Cho Jieun Park Jinyeong Chae Jihie Kim |
author_sort | Hojun Lee |
collection | DOAJ |
description | Transformer-based approaches have shown good results in image captioning tasks. However, current approaches have a limitation in generating text from global features of an entire image. Therefore, we propose novel methods for generating better image captioning as follows: (1) The Global-Local Visual Extractor (GLVE) to capture both global features and local features. (2) The Cross Encoder-Decoder Transformer (CEDT) for injecting multiple-level encoder features into the decoding process. GLVE extracts not only global visual features that can be obtained from an entire image, such as size of organ or bone structure, but also local visual features that can be generated from a local region, such as lesion area. Given an image, CEDT can create a detailed description of the overall features by injecting both low-level and high-level encoder outputs into the decoder. Each method contributes to performance improvement and generates a description such as organ size and bone structure. The proposed model was evaluated on the IU X-ray dataset and achieved better performance than the transformer-based baseline results, by 5.6% in BLEU score, by 0.56% in METEOR, and by 1.98% in ROUGE-L. |
first_indexed | 2024-03-09T21:06:45Z |
format | Article |
id | doaj.art-04b970dd6f314f269a9796220032b5e0 |
institution | Directory Open Access Journal |
issn | 1424-8220 |
language | English |
last_indexed | 2024-03-09T21:06:45Z |
publishDate | 2022-02-01 |
publisher | MDPI AG |
record_format | Article |
series | Sensors |
spelling | doaj.art-04b970dd6f314f269a9796220032b5e02023-11-23T21:59:22ZengMDPI AGSensors1424-82202022-02-01224142910.3390/s22041429Cross Encoder-Decoder Transformer with Global-Local Visual Extractor for Medical Image CaptioningHojun Lee0Hyunjun Cho1Jieun Park2Jinyeong Chae3Jihie Kim4Department of Computer Science and Engineering, Dongguk University, Seoul 04620, KoreaDepartment of Computer Science and Engineering, Dongguk University, Seoul 04620, KoreaDepartment of Computer Science and Engineering, Dongguk University, Seoul 04620, KoreaDepartment of Artificial Intelligence, Dongguk University, Seoul 04620, KoreaDepartment of Artificial Intelligence, Dongguk University, Seoul 04620, KoreaTransformer-based approaches have shown good results in image captioning tasks. However, current approaches have a limitation in generating text from global features of an entire image. Therefore, we propose novel methods for generating better image captioning as follows: (1) The Global-Local Visual Extractor (GLVE) to capture both global features and local features. (2) The Cross Encoder-Decoder Transformer (CEDT) for injecting multiple-level encoder features into the decoding process. GLVE extracts not only global visual features that can be obtained from an entire image, such as size of organ or bone structure, but also local visual features that can be generated from a local region, such as lesion area. Given an image, CEDT can create a detailed description of the overall features by injecting both low-level and high-level encoder outputs into the decoder. Each method contributes to performance improvement and generates a description such as organ size and bone structure. The proposed model was evaluated on the IU X-ray dataset and achieved better performance than the transformer-based baseline results, by 5.6% in BLEU score, by 0.56% in METEOR, and by 1.98% in ROUGE-L.https://www.mdpi.com/1424-8220/22/4/1429medical image captioningdeep learningtransformer |
spellingShingle | Hojun Lee Hyunjun Cho Jieun Park Jinyeong Chae Jihie Kim Cross Encoder-Decoder Transformer with Global-Local Visual Extractor for Medical Image Captioning Sensors medical image captioning deep learning transformer |
title | Cross Encoder-Decoder Transformer with Global-Local Visual Extractor for Medical Image Captioning |
title_full | Cross Encoder-Decoder Transformer with Global-Local Visual Extractor for Medical Image Captioning |
title_fullStr | Cross Encoder-Decoder Transformer with Global-Local Visual Extractor for Medical Image Captioning |
title_full_unstemmed | Cross Encoder-Decoder Transformer with Global-Local Visual Extractor for Medical Image Captioning |
title_short | Cross Encoder-Decoder Transformer with Global-Local Visual Extractor for Medical Image Captioning |
title_sort | cross encoder decoder transformer with global local visual extractor for medical image captioning |
topic | medical image captioning deep learning transformer |
url | https://www.mdpi.com/1424-8220/22/4/1429 |
work_keys_str_mv | AT hojunlee crossencoderdecodertransformerwithgloballocalvisualextractorformedicalimagecaptioning AT hyunjuncho crossencoderdecodertransformerwithgloballocalvisualextractorformedicalimagecaptioning AT jieunpark crossencoderdecodertransformerwithgloballocalvisualextractorformedicalimagecaptioning AT jinyeongchae crossencoderdecodertransformerwithgloballocalvisualextractorformedicalimagecaptioning AT jihiekim crossencoderdecodertransformerwithgloballocalvisualextractorformedicalimagecaptioning |