International Journal of Electrical and Computer Engineering (IJECE) Vol. 13, No. 6, December 2023, pp. 6293~6301 ISSN: 2088-8708, DOI: 10.11591/ijece.v13i6.pp6293-6301  6293 Journal homepage: http://ijece.iaescore.com Novel hybrid generative adversarial network for synthesizing image from sketch Pavithra Narasimha Murthy 1 , Sharath Kumar Yeliyur Hanumanthaiah 2 1 Maharaja Institute of Technology Mysuru, Visvesvaraya Technological University, Belagavi, India 2 Department of Information Science and Engineering, Maharaja Institute of Technology, Mysuru Visvesvaraya Technological University, Belagavi, India Article Info ABSTRACT Article history: Received Feb 6, 2023 Revised May 31, 2023 Accepted Jun 4, 2023 In the area of sketch-based image retrieval process, there is a potential difference between retrieving the match images from defined dataset and constructing the synthesized image. The former process is quite easier while the latter process requires more faster, accurate, and intellectual decision making by the processor. After reviewing open-end research problems from existing approaches, the proposed scheme introduces a computational framework of hybrid generative adversarial network (GAN) as a solution to address the identified research problem. The model takes the input of query image which is processed by generator module running 3 different deep learning modes of ResNet, MobileNet, and U-Net. The discriminator module processes the input of real images as well as output from generator. With a novel interactive communication between generator and discriminator, the proposed model offers optimal retrieval performance along with an inclusion of optimizer. The study outcome shows significant performance improvement. Keywords: Deep learning Generative adversarial network Hybrid model Image synthesis Sketch-based image retrieval This is an open access article under the CC BY-SA license. Corresponding Author: Pavithra Narasimha Murthy Maharaja Institute of Technology, Visvesvaraya Technological University Belagavi, India Email: pavithra.apr02@gmail.com 1. INTRODUCTION Search engines and tools has become one of the efficient ways to look for various forms of information both in standardized as well as in customized manner [1]. With constraints towards text-based search, image-based search option is one of the simplest and best option to look for information [2]. However, image-based search can be further more customized when the query image is given in the form of free hand drawing also known as sketch [3]. In this form of search technique, a sketch related to certain real object is drawn by user and it acts as an input to the system which looks for all matching images respective to the sketch. In this perspective, the relevance factor between the query images and real images in other sources are matched by the user using various shape-based information [4]. In such a process, generally the content-based information is not much prioritized, and this is one of the prime reasons for multiple outliers in the outcome of related images. The application of sketch-based image retrieval system is not merely about matching and extracting the related images, but the complexity is to generate the synthesized or reconstructed image by the machine [5]. At present, there are multiple approaches and techniques applied for this purpose with claimed beneficial characteristics. From the conventional scheme of information retrieval system, it is known that there are basically three types of standard approaches viz. i) text-based, ii) content-based, and iii) sketch based [6]. The existing research work towards this domain of problem is mainly emphasized towards improving various internal operations which has a direct impact towards retrieval process, and it is outcome. From this perspective, enough significance is given towards feature-based processing using