Publications

Coarse to Fine Bilinear Convolutional Neural Networks for Fine-grained Image Recognition

Published in (Pre-print), 2021

This paper proposes a simple yet novel way, “Coarse to Fine Bilinear Convolutional Neural Networks”, which preserves the coarse features while learning the finer features. Experiments have been performed on Hierarchical Grocery Store Image Datasets and Coarse to Fine Bilinear Convolutional Neural Networks had an accuracy of 85.33% which was better than Bilinear networks, which had an accuracy of 77.82%, and Linear networks, which had an accuracy of 72.11%.

Recommended citation: Saaketh Koundinya, You. (2021). "Paper Title Number 2." http://tommarvoloriddle.github.io/saaketh.github.io/files/Coarse-fine-BilinearCNN.pdf

Leakage Detection in Water Distribution Systems using Neural Networks

Published in Bachelor's Thesis, 2019

Used data from the local municipality to train a neural network for predicting leakages in the distribution systems. Applied this model to our college water distribution system.

Smart Bot for Handwritten Digit String Recognition

Published in 2022 International Conference for Advancement in Technology (ICONAT) Goa, India., -0022

The CNN model used in this paper determines the digit in the image provided real-time. The system plays the recorded audio and displays the output for the recognized digits in the given digit string.

Download here