我正在尝试将我的算法包含到 IEEE trans 的两列格式中的一列中。我的代码是：
\documentclass[conference]{IEEEtran}
\usepackage[algoruled,boxed,lined]{algorithm2e}
\usepackage{stfloats}
Computing objectness is very efficient and takes only about 4 sec. per image. This technique finds out some image windows like Figure \ref{objectnessfig}.\\
\begin{algorithm}
\SetAlgoLined
\begin{description}\itemsep0pt \parskip0pt \parsep0pt \vspace{.1cm}
\item[Input:] \(F, D, c\)
\item[Ouput:] \(Det\)
\item[Step 1:] \( \l = \left \{ w_{1},...,w_{F} \right \}, w_{i}\rightarrow D, \forall_{i} \)
\item[Step 2:] \( \l_{s} = \left \{ \left ( w_{1},sw_{1} \right ),...,\left ( w_{F},sw_{F} \right ) \right \}, sw_{i}= c\left ( w_{i} \right ), \forall_{i} \)
\item[Step 3:] \( \rho _{s} = NMS\left ( \l_{s} \right )=\left \{ \left ( w_{n1},sw_{n1} \right ),...,\left ( w_{np},sw_{np} \right ) \right \}\)
\item[Step 4:] \(\L=\left \{ w_{n1}^{lm},..., w_{np}^{lm} \right \}, w_{nj}^{lm} = max \left ( s_{w} \right )\)
\item[Step 5:] \(Det = NMS\left ( \L \right )\)
\end{description}
\caption{Using objectness for class-specific detectors.}
\end{algorithm}
The general scheme for using their objectness measure as
a location prior for object detectors is algorithm 1. The
algorithm inputs the class-specific confidence function \(c\) which
the detector employs to score a window.
They build an initial set \(\l\) of \(F = 1000\) windows multinomially sampled from the distribution \(D\) of windows scored by
their objectness measure (Multi-scale Saliency)\(MS\) +(Color Contrast)\(CC\) + (Superpixels Straddling)\(SS\) (step 1). They use \(c\) to
score each window in \(\l\) (step 2). They then run the non-maxima
suppression. This results in a set \(\rho_{s}\) of promising
windows (step 3). For every window \(w_{p} \epsilon \rho_{s}\), they iteratively
move to the local maximum of \(c\) in its neighborhood \(V_{w
p}\),
resulting in window \(w_{p}^{lm}\) (step 4). Finally, they run \(NMS\) on the
local maxima windows \(\L\) and obtain detections \(Det\) (step 5).
In order to use this algorithm one has to specify a window
scoring function \(c\), which is specific to a particular detector
and object class, and a window neighborhood.
但问题是我得到了以下结果。
我不知道为什么算法旁边的空间是空的。如果我注释掉算法，它就没问题了。到底是什么问题？？
编辑：当我使用 \Blindtext 时，算法表放置得很好，但是当我提供我的文本时，它就会显示上述问题。
我也遇到了同样的问题，我想让表格保留在一列中。
。
但是当我使用 \Blindtext 时就没问题了。
下面是我用于表格的代码：
\begin{table}
\centering
\caption{Result for the sample image for Selective Search}
\label{ressel}
\begin{tabularx}{\linewidth}{|*{5}{X|}}
\hline
Labels & Label-1 & Label-2 & Label-3 & Label-4 \\ \hline
Top-1 & bird & cat & frog & dog \\ \hline
Top-2 & cat & bird & dog & frog \\ \hline
Cumulative & cat & frog & bird & dog \\ \hline
Total Score & cat & frog & bird & dog \\ \hline
\end{tabularx}
\end{table}
答案1
我正在使用\usepackage{stfloats}。删除包解决了我的问题。