This repository provides the code and trained model from our paper: M. Tzelepi, V. Mezaris, "Online Anchor-based Training for Image Classification Tasks", Proc. 2024 IEEE Int. Conf. on Image Processing (ICIP 2024), Abu Dhabi, UAE, pp. 1099-1105, Oct. 2024. DOI:10.1109/ICIP51287.2024.10648148.
In this work, we improve the performance of a deep learning model towards image classification tasks by proposing an anchor-based training methodology, named Online Anchor-based Training (OAT). The OAT method, guided by the insights provided in the anchor-based object detection methodologies, instead of learning directly the class labels, proposes to train a model to learn percentage changes of the class labels with respect to defined anchors. We define as anchors the batch centers at the output of the model. Then, during the test phase, the predictions are converted back to the original class label space, and the performance is evaluated. The effectiveness of the OAT method is validated on four datasets.
- OAT training and evaluation
python3 oat_efficientnet.py --nof_epochs 90 --l_r 0.001 --batch_size 32 --nof_classes 6 --pretrained True --weights_path /path/to/efficient10.pth- Baseline training and evaluation
python3 baseline_efficientnet.py --nof_epochs 100 --l_r 0.001 --batch_size 32 --nof_classes 6Evaluation results in terms of test accuracy (%) of the OAT method (with and without pretraining) against baseline, using the EfficientNet-b1 model (5 runs).
| Method | BAR |
|---|---|
| Baseline | 67.278 |
| OAT w/o pretraining | 68.593 |
| OAT w pretraining | 69.174 |
This project has been funded by the European Union as part of the Horizon Europe Framework Program, under grant agreement 101070109 (TransMIXR).
This code is provided for academic, non-commercial use only. Please also check for any restrictions applied in the code parts and datasets used here from other sources. For the materials not covered by any such restrictions, redistribution and use in source and binary forms, with or without modification, are permitted for academic non-commercial use provided that the following conditions are met:
Redistributions of source code must retain the above copyright notice, this list of conditions and the following disclaimer. Redistributions in binary form must reproduce the above copyright notice, this list of conditions and the following disclaimer in the documentation provided with the distribution.
This software is provided by the authors "as is" and any express or implied warranties, including, but not limited to, the implied warranties of merchantability and fitness for a particular purpose are disclaimed. In no event shall the authors be liable for any direct, indirect, incidental, special, exemplary, or consequential damages (including, but not limited to, procurement of substitute goods or services; loss of use, data, or profits; or business interruption) however caused and on any theory of liability, whether in contract, strict liability, or tort (including negligence or otherwise) arising in any way out of the use of this software, even if advised of the possibility of such damage.
If you find our method useful in your work or you use some materials provided in this repo, please cite the following publication where our method and materials were presented:
DOI: https://doi.org/10.1109/ICIP51287.2024.10648148
Pre-print: http://arxiv.org/abs/2406.12662
@inproceedings{tzelepi2024online,
title={Online Anchor-based Training for Image Classification Tasks},
author={Tzelepi, Maria and Mezaris, Vasileios},
year={2024},
booktitle={Proc. 2024 IEEE International Conference on Image Processing (ICIP 2024)}
}