Publication Template

Title	GeoRTNet: Geometric Perception based Efficient Text Recognition
Date	March 2021
Venue	Journal of Machine Learning Research (JMLR)
Status	In Review
Authors	John Doe, Jane Smith, Mary Johnson
keyword	Neural Style Transfer, Conditional Adversarial Networks, GAN, Deep Learning
Code	Paper

Abstract

This project aimed to develop a mobile app that uses deep learning techniques, specifically Convolutional Neural Networks (CNNs), to identify objects in images captured by the user's smartphone camera.

Key Contributions

pre-trained CNN models like MobileNetV2
fine-tuned them on a custom dataset
optimized the model for real-time performance on mobile devices using TensorFlow Lite

Lessons Learned

Proficiency in deep learning
real-time image processing
model optimization