Saved in:
Bibliographic Details
Main Author: Huang, Feiyang
Format: Preprint
Published: 2024
Subjects:
Online Access:https://arxiv.org/abs/2411.07265
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866916693412937728
author Huang, Feiyang
author_facet Huang, Feiyang
contents This paper presents ViTOC (Vision Transformer and Object-aware Captioner), a novel vision-language model for image captioning that addresses the challenges of accuracy and diversity in generated descriptions. Unlike conventional approaches, ViTOC employs a dual-path architecture based on Vision Transformer and object detector, effectively fusing global visual features and local object information through learnable vectors. The model introduces an innovative object-aware prompting strategy that significantly enhances its capability in handling long-tail data. Experiments on the standard COCO dataset demonstrate that ViTOC outperforms baseline models across all evaluation metrics. Additionally, we propose a reference-free evaluation method based on CLIP to further validate the model's effectiveness. By utilizing pretrained visual model parameters, ViTOC achieves efficient end-to-end training.
format Preprint
id arxiv_https___arxiv_org_abs_2411_07265
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle ViTOC: Vision Transformer and Object-aware Captioner
Huang, Feiyang
Computer Vision and Pattern Recognition
This paper presents ViTOC (Vision Transformer and Object-aware Captioner), a novel vision-language model for image captioning that addresses the challenges of accuracy and diversity in generated descriptions. Unlike conventional approaches, ViTOC employs a dual-path architecture based on Vision Transformer and object detector, effectively fusing global visual features and local object information through learnable vectors. The model introduces an innovative object-aware prompting strategy that significantly enhances its capability in handling long-tail data. Experiments on the standard COCO dataset demonstrate that ViTOC outperforms baseline models across all evaluation metrics. Additionally, we propose a reference-free evaluation method based on CLIP to further validate the model's effectiveness. By utilizing pretrained visual model parameters, ViTOC achieves efficient end-to-end training.
title ViTOC: Vision Transformer and Object-aware Captioner
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2411.07265