Reliable Deep Learning for Small-Scale Classifications: Experiments on Real-World Image Datasets from Bangladesh

TL;DR

This paper evaluates a compact CNN architecture on five real-world small-scale image datasets from Bangladesh, demonstrating high accuracy, efficiency, and robustness in diverse classification scenarios.

Contribution

It provides empirical evidence that streamlined CNNs can effectively handle small datasets with diverse real-world images, reducing overfitting and computational costs.

Findings

High classification accuracy across datasets

Efficient convergence and low computational overhead

Robust generalization demonstrated by saliency analysis

Abstract

Convolutional neural networks (CNNs) have achieved state-of-the-art performance in image recognition tasks but often involve complex architectures that may overfit on small datasets. In this study, we evaluate a compact CNN across five publicly available, real-world image datasets from Bangladesh, including urban encroachment, vehicle detection, road damage, and agricultural crops. The network demonstrates high classification accuracy, efficient convergence, and low computational overhead. Quantitative metrics and saliency analyses indicate that the model effectively captures discriminative features and generalizes robustly across diverse scenarios, highlighting the suitability of streamlined CNN architectures for small-class image classification tasks.