Abstract
Convolutional Neural Networks (CNNs) are a standard approach for visual recognition due to their capacity to learn hierarchical representations from raw pixels. In practice, practitioners often choose among (i) training a compact custom CNN from scratch, (ii) using a large pre-trained CNN as a fixed feature extractor, and (iii) performing transfer learning via partial or full fine-tuning of a pre-trained backbone. This report presents a controlled comparison of these three paradigms across five real-world image classification datasets spanning road-surface defect recognition, agricultural variety identification, fruit/leaf disease recognition, pedestrian walkway encroachment recognition, and unauthorized vehicle recognition. Models are evaluated using accuracy and macro F1-score, complemented by efficiency metrics including training time per epoch and parameter counts. The results show that transfer learning consistently yields the strongest predictive performance, while the custom CNN provides an attractive efficiency--accuracy trade-off, especially when compute and memory budgets are constrained.
Abstract (translated)
卷积神经网络（CNN）由于其从原始像素中学习分层表示的能力，成为视觉识别的标准方法。在实践中，从业人员通常会在这三种方式之间进行选择：(i) 从头开始训练紧凑型自定义的 CNN；(ii) 使用大型预训练的 CNN 作为固定的特征提取器；以及 (iii) 通过部分或完全微调预训练的基础模型来进行迁移学习。本报告对这三种范式进行了严格的比较，涵盖五个真实世界图像分类数据集，这些数据集涉及道路表面缺陷识别、农业品种识别、水果/叶片疾病识别、人行道侵入识别和未经授权的车辆识别。评估使用了准确率和宏F1分数指标，并辅以效率度量，包括每个训练周期的时间以及参数数量。结果表明，在预测性能方面，迁移学习始终表现出色；而定制 CNN 在计算资源（如计算能力和内存）受限的情况下提供了高效性和准确性之间的良好权衡。
URL
https://arxiv.org/abs/2601.02246
