From 2a95b44d8a5df802fe0a144d8a0d7fd8d041142d Mon Sep 17 00:00:00 2001 From: Jim Martens Date: Thu, 7 Mar 2019 16:43:52 +0100 Subject: [PATCH] Unified headlines to be in title case and changed open-set to open set Signed-off-by: Jim Martens --- body_expose.tex | 20 ++++++++++---------- 1 file changed, 10 insertions(+), 10 deletions(-) diff --git a/body_expose.tex b/body_expose.tex index be03d7f..90e0098 100644 --- a/body_expose.tex +++ b/body_expose.tex @@ -50,9 +50,9 @@ describes tasks where the network is supposed to identify the class of any given input. In this thesis, I will focus on classification. -\subsection*{Object detection in open-set conditions} +\subsection*{Object Detection in Open Set Conditions} -More specifically, I will look at object detection in the open-set +More specifically, I will look at object detection in the open set conditions. In non-technical words this effectively describes the kind of situation you encounter with CCTV cameras or robots outside of a laboratory. Both use cameras that record @@ -108,7 +108,7 @@ auto-encoder, a novelty score is calculated. A low novelty score signals a known object. The opposite is true for a high novelty score. -\subsection*{Research question} +\subsection*{Research Question} Given these two approaches to solve the explanation task of above, it comes down to performance. At the end of the day the best @@ -131,14 +131,14 @@ being equal. With respect to auto-encoders a recent implementation of an adversarial auto-encoder\cite{Pidhorskyi2018} will be used. \paragraph{Hypothesis} Novelty detection using auto-encoders -delivers similar or better object detection performance under open-set +delivers similar or better object detection performance under open set conditions while being less computationally expensive compared to dropout sampling. \paragraph{Contribution} The contribution of this thesis is a comparison between dropout sampling and auto-encoding with respect to the overall performance -of both for object detection in the open-set conditions using +of both for object detection in the open set conditions using the SSD network for object detection and the SceneNet RGB-D data set with MS COCO classes. @@ -175,7 +175,7 @@ and intractable problem of averaging over all weights in the network is replaced with an optimisation task, where the parameters of the simple distribution are optimised over\cite{Kendall2017}. -\subsubsection*{Dropout variational inference} +\subsubsection*{Dropout Variational Inference} Kendall and Gal\cite{Kendall2017} showed an approximation for classfication and recognition tasks. Dropout variational inference @@ -199,7 +199,7 @@ With this dropout sampling technique \(n\) model weights of the network with respect to the classification is given by the entropy \(H(\mathbf{q}) = - \sum_i q_i \cdot \log q_i\). -\subsubsection*{Dropout sampling for object detection} +\subsubsection*{Dropout Sampling for Object Detection} Miller et al\cite{Miller2018} apply the dropout sampling to object detection. In that case \(\mathbf{W}\) represents the @@ -230,7 +230,7 @@ distribution means that no class is more likely than another, which is a perfect example of maximum uncertainty. Conversely, if one class has a very high probability the entropy will be low. -In open-set conditions it can be expected that falsely generated +In open set conditions it can be expected that falsely generated detections for unknown object classes have a higher label uncertainty. A treshold on the entropy \(H(\mathbf{q}_i)\) can then be used to identify and reject these false positive cases. @@ -370,7 +370,7 @@ the one of dropout sampling or if the object detection performance is worse. \paragraph{Hypothesis} Novelty detection using auto-encoders -delivers similar or better object detection performance under open-set +delivers similar or better object detection performance under open set conditions while being less computationally expensive compared to dropout sampling.\\ @@ -389,7 +389,7 @@ the thesis: Summarising, the scientific contribution is a comparison between dropout sampling and GPND with respect to both object detection -performance and computational performance under open-set conditions +performance and computational performance under open set conditions using the SceneNet RGB-D data set with the MS COCO classes as "known" object classes.