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

## TL;DR

This paper presents ELF, a CNN-based feature detector that extracts keypoint locations from gradient information within pre-trained networks, requiring no additional training and achieving competitive performance.

## Contribution

ELF introduces a novel method to derive feature locations from pre-trained CNNs without supervised training or fine-tuning, leveraging gradient-based saliency maps.

## Key findings

- Achieves similar repeatability and matching scores as traditional detectors.
- Demonstrates robustness to illumination and viewpoint changes.
- Operates without additional training or fine-tuning.

## Abstract

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.