diff options
author | Vasil Zlatanov <v@skozl.com> | 2019-03-22 23:10:59 +0000 |
---|---|---|
committer | Vasil Zlatanov <v@skozl.com> | 2019-03-22 23:10:59 +0000 |
commit | 0e9a762665554b0ff0d13f088366fd09a25a69a2 (patch) | |
tree | 177220a4aaa35e2bcac1b6adc7f13d8ff14c2d73 /report/template.latex | |
parent | 5245b5ec835175d390b78f5b201e7ca5fdaaab9b (diff) | |
download | e3-deep-master.tar.gz e3-deep-master.tar.bz2 e3-deep-master.zip |
Diffstat (limited to 'report/template.latex')
-rw-r--r-- | report/template.latex | 6 |
1 files changed, 5 insertions, 1 deletions
diff --git a/report/template.latex b/report/template.latex index 232f8b3..8d27aef 100644 --- a/report/template.latex +++ b/report/template.latex @@ -47,7 +47,7 @@ %%%%%%%%% ABSTRACT \begin{abstract} - Abstract - In this paper we investigate triplet loss based methods for verification, matching and retrieval of the HPatches dataset. We explore a system of two models, one for denoising of the patches and one for generation of descriptors. We are able to show that a model trained with online hard patch mining, while more difficult, is able to achieve improved performance over a classical Siamese triplet network. + Abstract - In this paper we investigate triplet loss based methods for verification, matching and retrieval of the HPatches dataset. We explore a system of two models - one for denoising of the patches and one for generation of descriptors. We show that a model utilising online hard mining, while more difficult, is able to achieve improved performance over a classical Siamese triplet network. \end{abstract} \providecommand{\tightlist}{% @@ -77,6 +77,10 @@ $endif$ \newcommand{\nnfn}{f_\theta} \newcommand{\norm}[1]{\left\lVert#1\right\rVert} % Thanks http://tex.stackexchange.com/a/107190 +\makeatletter +\def\fps@figure{h} +\makeatother + $body$ $if(natbib)$ |