Analysis of Mild Overfitting, Class-Level Performance, and Future Directions in AI-Driven Fashion
Preprint
- 27 May 2026
- preprint
- Published by MDPI AG in Preprints.org
Abstract
The given research paper describes a CNN model of classifying images belonging to more than two classes on the Fashion-MNIST data. The model performed a test accuracy of 92.44% and test loss of 0.2533 the greatest accuracy as compared to similar studies with similar architectures. The architecture has three convolutional-pooling blocks, a dense layer with dropout regularization (0.3), and a softmax output layer. The analysis of training and validation curves demonstrates mild overfitting of the later epochs, and the validation loss starts growing even though the training loss continues to decrease. In-depth analysis using confusion matrix and classification report identifies certain patterns of misclassification between visually similar categories. The paper also discusses implications on batch normalization, data augmentation as well as Vision Transformer architecture.Keywords
All Related Versions
This publication has 37 references indexed in Scilit:
- Digitizing historical Kannada: An OCR approach for Yakshagana scripts based on glyph-to-Unicode dictionariesDigital Applications in Archaeology and Cultural Heritage, 2026
- AeroCOPDNet: A deep learning framework for COPD detection from lung soundsBiomedical Signal Processing and Control, 2026
- AI-Based Detection of Postural Anomalies for Sport Medicine and Physiotherapy: Comparative Deep Learning and Clinical Thresholding Approaches2026
- Investigating the Influence of Loss Functions on the Performance and Interpretability of Machine Learning Models2025
- A Comprehensive Survey on Emerging AI Technologies for 6G Communications: Research Direction, Trends, Challenges, and OpportunitiesInternational Journal of Intelligent Networks, 2025
- Experimental Investigation on Mechanical Characterization of Epoxy-E-Glass Fiber-Particulate Reinforced Hybrid CompositesACS Omega, 2024
- Superposition-enhanced quantum neural network for multi-class image classificationChinese Journal of Physics, 2024
- Image classification of root-trimmed garlic using multi-label and multi-class classification with deep convolutional neural networkPostharvest Biology and Technology, 2022
- New techniques for efficiently k-NN algorithm for brain tumor detectionMultimedia Tools and Applications, 2022
- Multi-class Segmentation of Organ at Risk from Abdominal CT Images: A Deep Learning Approach2021