INTERNATIONAL JOURNAL OF SCIENTIFIC DEVELOPMENT AND RESEARCH International Peer Reviewed & Refereed Journals, Open Access Journal ISSN Approved Journal No: 2455-2631 | Impact factor: 8.15 | ESTD Year: 2016
open access , Peer-reviewed, and Refereed Journals, Impact factor 8.15
ENHANCING TEXTUAL DESCRIPTION USING DEEP LEARNING WITH RC-GAN GENERATION
Authors Name:
A.Srujan Reddy
, N. Venkata Mahesh Kumar Reddy , N.Bharath , S. Raaga Sindhu , R. Jagadeeswari
Unique Id:
IJSDR2404110
Published In:
Volume 9 Issue 4, April-2024
Abstract:
One method employed to create pixels representations corresponding to verbal descriptions is known as "text-to-image generation," impacting a broad spectrum of applications and research domains, including exploration, design, computer-aided drafting, image restoration, labelling, and portraiture. The primary challenge lies in consistently generating realistic photographs under specified conditions. Existing algorithms for script-to-image conversion often fail to accurately align images with the accompanying text descriptions. In our investigation, we addressed this challenge by devising a deep learning architecture named the recurrent convolutional generative adversarial network (RC-GAN) specifically fine-tuned for producing semantically concise snapshots. RC-GAN effectively translates conceptual visual elements from linguistic cues to pixel-level representations, bridging the gap between advancements in text and image modelling. To train our proposed model, we utilised various datasets, employing metrics such as inception score and peak signalto-noise ratio (PSNR) to analyse its performance. Empirical findings showcase an inception required score and a PSNR required measurement, indicating that our model can produce more realistic images from verbal descriptions. Our future endeavour involves training the proposed system on various datasets to further improve its capabilities.
Keywords:
Deep Learning (DL), Convolutional Neural Network (CNN), Recurrent Neural Network (RNN), Generation of Image, Generative Adversarial Networks (GAN).
Cite Article:
"ENHANCING TEXTUAL DESCRIPTION USING DEEP LEARNING WITH RC-GAN GENERATION ", International Journal of Science & Engineering Development Research (www.ijsdr.org), ISSN:2455-2631, Vol.9, Issue 4, page no.787 - 792, April-2024, Available :http://www.ijsdr.org/papers/IJSDR2404110.pdf
Downloads:
000338172
Publication Details:
Published Paper ID: IJSDR2404110
Registration ID:210836
Published In: Volume 9 Issue 4, April-2024
DOI (Digital Object Identifier):
Page No: 787 - 792
Publisher: IJSDR | www.ijsdr.org
ISSN Number: 2455-2631
Facebook Twitter Instagram LinkedIn