# Good Practice in CNN Feature Transfer

arXiv: https://arxiv.org/abs/1604.00133 · HTML: https://mlsanity.com/paper/1604.00133/

## Key facts

- Cited by 12 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 1 Apr 2016.

## Built on by (most cited first, 12 of 12)

- [Perception Encoder: The best visual embeddings are not at the output of the network](https://mlsanity.com/paper/2504.13181/) (2025)
- [Deep Learning for Instance Retrieval: A Survey](https://mlsanity.com/paper/2101.11282/) (2021)
- [StyLIP: Multi-Scale Style-Conditioned Prompt Learning for CLIP-based Domain Generalization](https://mlsanity.com/paper/2302.09251/) (2023)
- [Investigating the Role of Image Retrieval for Visual Localization -- An exhaustive benchmark](https://mlsanity.com/paper/2205.15761/) (2022)
- [Random Boxes Are Open-world Object Detectors](https://mlsanity.com/paper/2307.08249/) (2023)
- [Real-time Aerial Detection and Reasoning on Embedded-UAVs](https://mlsanity.com/paper/2305.12414/) (2023)
- [Few-shot learning for COVID-19 Chest X-Ray Classification with Imbalanced Data: An Inter vs. Intra Domain Study](https://mlsanity.com/paper/2401.10129/) (2024)
- [Multi-Label Logo Recognition and Retrieval based on Weighted Fusion of Neural Features](https://mlsanity.com/paper/2205.05419/) (2022)
- [Improving thermal state preparation of Sachdev-Ye-Kitaev model with reinforcement learning on quantum hardware](https://mlsanity.com/paper/2501.11454/) (2025)
- [Web Table Classification based on Visual Features](https://mlsanity.com/paper/2103.05110/) (2021)
- [MARs: Multi-view Attention Regularizations for Patch-based Feature Recognition of Space Terrain](https://mlsanity.com/paper/2410.05182/) (2024)
- [A Dual-stage Prompt-driven Privacy-preserving Paradigm for Person Re-Identification](https://mlsanity.com/paper/2511.05092/) (2025)

## Abstract (arXiv)

The objective of this paper is the effective transfer of the Convolutional Neural Network (CNN) feature in image search and classification. Systematically, we study three facts in CNN transfer. 1) We demonstrate the advantage of using images with a properly large size as input to CNN instead of the conventionally resized one. 2) We benchmark the performance of different CNN layers improved by average/max pooling on the feature maps. Our observation suggests that the Conv5 feature yields very competitive accuracy under such pooling step. 3) We find that the simple combination of pooled features extracted across various CNN layers is effective in collecting evidences from both low and high level descriptors. Following these good practices, we are capable of improving the state of the art on a number of benchmarks to a large margin.