Evolving CNN Architectures: From Custom Designs to Deep Residual Models for Diverse Image Classification and Detection Tasks

TL;DR

This study compares custom CNN architectures with pretrained models across diverse image tasks, highlighting how architectural choices affect performance and demonstrating the adaptability of CNNs in object detection scenarios.

Contribution

It provides a systematic analysis of CNN design factors and their impact on various image classification and detection tasks, offering practical guidance for model selection.

Findings

Deeper CNNs improve fine-grained recognition accuracy.

Pretrained models excel in binary classification tasks.

Custom architectures can be adapted for object detection in real-world scenes.

Abstract

This paper presents a comparative study of a custom convolutional neural network (CNN) architecture against widely used pretrained and transfer learning CNN models across five real-world image datasets. The datasets span binary classification, fine-grained multiclass recognition, and object detection scenarios. We analyze how architectural factors, such as network depth, residual connections, and feature extraction strategies, influence classification and localization performance. The results show that deeper CNN architectures provide substantial performance gains on fine-grained multiclass datasets, while lightweight pretrained and transfer learning models remain highly effective for simpler binary classification tasks. Additionally, we extend the proposed architecture to an object detection setting, demonstrating its adaptability in identifying unauthorized auto-rickshaws in real-world traffic scenes. Building upon a systematic analysis of custom CNN architectures alongside pretrained and transfer learning models, this study provides practical guidance for selecting suitable network designs based on task complexity and resource constraints.