Learning Covariant Feature Detectors

Karel Lenc and Andrea Vedaldi

A source code and network models for a translation covariant detector presented in "Learning Covariant Feature Detectors". Written in MATLAB using the MatConvNet library.

This code depends on VLFeat and MatConvNet library. The script setup.m attempts to download and install those if the libraries are not in the MATLAB path. The script setup.m also downloads the model files.

Currently contains the following models:

• ./nets/detnet_s1.mat Densely evaluation model DetNet-S
• ./nets/detnet_s2.mat DetNet-S evaluated with stride 2
• ./nets/detnet_s4.mat DetNet-S evalyated with stride 4

An example how to run the detector is shown in example.m which produces the following figure.