Neural Style Transfer using Telegram Bot
Venkata Bala Annapurna P, T. Hemanth Kumar, J.V V Sri Surya, T. Abhiram, K. Ramu, Thammali Gangadhar · 2025
Neural Style Transfer (NST) allows artistic image conversion through content and style fusion, but existing implementations are typically resource-intensive, time-consuming, and technical. It thus remains inaccessible to non-technical users. To address these concerns, this project introduces a lightweight Telegram bot that performs real-time style transfer using a pre-trained Adaptive Instance Normalization (AdaIN) model. AdaIN allows rapid high-quality output with reduced computation and hence can be used on a daily basis. The bot provides a simple user interface to translate artistic styles onto their photos without installing advanced software or requiring technical expertise. The approach is to simplify NST and make it more convenient and rapid to use.