# ELF: Embedded Localisation of Features in pre-trained CNN

arXiv: https://arxiv.org/abs/1907.03261 · HTML: https://mlsanity.com/paper/1907.03261/

## Key facts

- Cited by 9 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 7 Jul 2019.

## Built on by (most cited first, 9 of 9)

- [Deep Lucas-Kanade Homography for Multimodal Image Alignment](https://mlsanity.com/paper/2104.11693/) (2021)
- [BALF: Simple and Efficient Blur Aware Local Feature Detector](https://mlsanity.com/paper/2211.14731/) (2022)
- [Neural Reprojection Error: Merging Feature Learning and Camera Pose Estimation](https://mlsanity.com/paper/2103.07153/) (2021)
- [Wide-Baseline Multi-Camera Calibration using Person Re-Identification](https://mlsanity.com/paper/2104.08568/) (2021)
- [A Critical Analysis of Image-based Camera Pose Estimation Techniques](https://mlsanity.com/paper/2201.05816/) (2022)
- [Deep Learning Reforms Image Matching: A Survey and Outlook](https://mlsanity.com/paper/2506.04619/) (2025)
- [Multi-Person 3D Pose Estimation from Multi-View Uncalibrated Depth Cameras](https://mlsanity.com/paper/2401.15616/) (2024)
- [Image Feature Information Extraction for Interest Point Detection: A Review](https://mlsanity.com/paper/2106.07929/) (2021)
- [Looking Beyond Corners: Contrastive Learning of Visual Representations for Keypoint Detection and Description Extraction](https://mlsanity.com/paper/2112.12002/) (2021)

## Abstract (arXiv)

This paper introduces a novel feature detector based only on information embedded inside a CNN trained on standard tasks (e.g. classification). While previous works already show that the features of a trained CNN are suitable descriptors, we show here how to extract the feature locations from the network to build a detector. This information is computed from the gradient of the feature map with respect to the input image. This provides a saliency map with local maxima on relevant keypoint locations. Contrary to recent CNN-based detectors, this method requires neither supervised training nor finetuning. We evaluate how repeatable and how matchable the detected keypoints are with the repeatability and matching scores. Matchability is measured with a simple descriptor introduced for the sake of the evaluation. This novel detector reaches similar performances on the standard evaluation HPatches dataset, as well as comparable robustness against illumination and viewpoint changes on Webcam and photo-tourism images. These results show that a CNN trained on a standard task embeds feature location information that is as relevant as when the CNN is specifically trained for feature detection.