SS SRIVIDYA
SRINIVAS

Research

Published work.

Research exploring how visual geometry, dataset construction, and image preprocessing influence the performance of convolutional neural networks.

Testing accuracy comparison for CNN image classification experiments
Testing accuracy across object classes and CNN configurations.
Comparison between original and proportionally resized images
Original and proportionally resized image datasets.
CNN training and validation accuracy and loss over training epochs
Training and validation accuracy and loss across epochs.

IEEE ICCECE 2025

Understanding the Roles of Geometric Forms and Proportions in CNN-Based Image Classification

Co-authored with Prof. Anindya Deb | Indian Institute of Science

This research investigates how geometric variation, dataset size, CNN architecture, and proportional image resizing affect image-classification performance. The work compares natural and human-made objects to understand how differences in visual structure influence model learning and generalization.

Research question

How do object geometry, dataset composition, image proportions, and CNN design choices affect classification accuracy across natural and human-made objects?

Approach

Two custom image datasets were developed. The first contained cars, trucks, and roses, while the second contained lionesses and kittens. CNN performance was evaluated across different dataset sizes, convolutional layers, training epochs, batch sizes, and learning rates. The study also tested proportional resizing based on the approximate real-world size relationship between a lioness and a kitten.

Key results

98.25%

Training accuracy achieved on the cars, trucks, and roses dataset.

53.33% to 70%

Improvement in lioness and kitten testing accuracy after proportional resizing.

Geometry

Natural forms such as roses were more difficult to classify than visually structured human-made objects.

Technical focus

The experiments examined the relationship between object geometry and CNN behavior, including how scale, proportional preprocessing, network depth, and dataset composition influence classification accuracy.

Python CNNs Deep Learning Image Classification ImageDataGenerator Adam Optimizer Data Preprocessing
Read full paper ↗

DOI: 10.1109/ICCECE61355.2025.10941645