Hybrid Architecture using CNN and LSTM for Image Captioning in Hindi Language

Poddar, Ayush Kumar; Rani, Rajneesh

doi:10.1016/j.procs.2023.01.049

Cited by 20 publications

(2 citation statements)

References 8 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The output is subsequently human-evaluated using the Linguistic Data Consortium's correctness criteria, which are also utilized in NIST projects. A CNN-LSTM neural network model [81] for generating Hindi captions for photographs is presented in this paper. The model's effectiveness was assessed, and the results showed a 34.64 percent and 29.13 percent rise in BLEU scores when compared to previous work.…”

Section: Image Captioning For the Hindi Languagementioning

confidence: 99%

A Comprehensive Survey on Image Captioning for Indian Languages: Techniques, Datasets, and Challenges

Jayaswal,

Rani,

Kaur

2023

Preprint

View full text Add to dashboard Cite

In image captioning, we generate visual descriptions from an image. Image Cap-tioning requires identifying the key entity, feature, and association in an image. There is also a requirement to generate captions that are syntactically and semantically correct. The process of image captioning requires computer vision and natural language processing. In the past few decades, a substantial attempt has been made to generate the caption for images. In this survey article, we are going to present an extensive survey on image captioning for Indian Languages. To summarize recent research work in image captioning, first, we briefly review the traditional approach to image captioning depending on template and retrieval. Further deep-learning approaches for image captioning are concentrated which are classified as encoder-decoder architecture, attention-based approach, and transformer architecture. Our main focus in this survey is based on image cap-tioning techniques for Indian languages like Hindi, Bengali Assamese, etc. After that, we analyze the state-of-the-art approach on the most widely dataset i.e. MS COCO dataset with their strengths, limitations, and performance metrics i.e. BLEU, ROUGE, METEOR, CIDEr, SPICE. At last, we explore discussion on open challenges and future direction in the field of image captioning.

show abstract

Section: Image Captioning For the Hindi Languagementioning

confidence: 99%

A Comprehensive Survey on Image Captioning for Indian Languages: Techniques, Datasets, and Challenges

Jayaswal,

Rani,

Kaur

2023

Preprint

View full text Add to dashboard Cite

show abstract

“…However, encoder-decoder-based approaches are not capable of analyzing the images over time and considering the spatial prospects of images that are pertinent to the image description (alternatively, creating descriptions for the entire scene). Recently, to conquer the previous restrictions, attention mechanisms have been utilized for mapping text captioning to various areas of the image [8]. In this paper, the principal contributions are specified as follows:…”

Section: Introductionmentioning

confidence: 99%

Image to Text Description Approach based on Deep Learning Models

Hameed Arif

2024

BAJEST

View full text Add to dashboard Cite

The image-to-text description can be indicated by creating captions for images that comply with human language perception. Nowadays, with the speedy progress of deep learning models, image-to-text description (or image captioning) has an expanding consideration by numerous researchers in diverse artificial intelligence relevant applications. In general, accurately getting the semantic information of the principal objects in the images and captioning the association among them represents a crucial issue in this field. In this paper, an image-to-text description approach based on Inception-ResNetV2-LSTM with an attention technique is proposed for effective textual descriptions of images. In this proposed approach, Inception-ResNetV2 is exploited to extract essential features, and the integration of LSTM with the attention technique is implemented as a sentence-creation model in such a way that the learning could be concentrated on specific portions within the images, hence enhancing the performance of image-to-text description approach. In terms of the Meteor and BLEU (1-4) measurements, the proposed approach outperformed other state-of-the-art approaches with 0.787 and (0.977, 0.964, 0.886, and 0.759), respectively

show abstract