Merge branch 'main' of https://github.com/Sproc01/8BallPool

Report/main.tex

@@ -105,10 +105,10 @@
 % directly follow a colon or long dash
 \title{8BallPool video analysis}
 \author{
-    %Authors
-    Michele Sprocatti\textsuperscript{\rm 1}, %\equalcontrib,
-    Alberto Pasqualetto\textsuperscript{\rm 2}, %\equalcontrib,
-    Michela Schibuola\textsuperscript{\rm 3}\\
+	%Authors
+	Michele Sprocatti\textsuperscript{\rm 1}, %\equalcontrib,
+	Alberto Pasqualetto\textsuperscript{\rm 2}, %\equalcontrib,
+	Michela Schibuola\textsuperscript{\rm 3}\\
 }
 
 \begin{document}

Report/section/elements.tex

@@ -1,5 +1,4 @@
 \section{Elements of our project}	% TODO change name
-
 % TODO ThE FOLLOWING ARE AGAIN SECTIONS -> FIX
 \input{section/executables}
 \input{section/tableDetection}

Report/section/executables.tex

@@ -6,3 +6,5 @@ \section{Executables}
 	\item \texttt{ShowSegmentationColored}: is an helper which has been used to show the ground truth of the segmentation of a particular frame using human-readable colors and it was also used as a test for the code that computes the metrics because it computes the performance of the ground truth on itself.
 	\item \texttt{ComputePerformance}: is used to compute the performance across the dataset so the mAP and the mIoU.
 \end{itemize}
+
+% TODO cmd parameters

Report/section/introduction.tex

@@ -1,3 +1,2 @@
 \section{Introduction}
-
-This is the report of the final project for the Computer Vision course, which goal is to develop a computer vision system for analyzing video footage of various “Eight Ball” billiard game events.
+This is the report of the final project for the Computer Vision course, which goal is to develop a computer vision system for analyzing video footage of various “Eight Ball” billiard game events.

Report/section/results.tex

@@ -1,5 +1,4 @@
 \section{Results}
-
 Table detection exhibits very high accuracy across the dataset, in particular for each initial frame of each video 4 corner points are always identified, the assumption that we made is that the camera does not move during a single clip so, once the table is detected in the first frame, we can use that information for all the other frames in the same video.
 
 In contrast, ball detection is influenced by k-means clustering. To achieve consistent and satisfactory results, a fixed random seed is incorporated into the code. %TODO scrivere perchè

Report/section/tableDetection.tex

@@ -1,5 +1,4 @@
 \subsection{Table detection}
-
 A mask-based approach was implemented for table detection, exploiting the fact that in billiard games' footage, the table should be always located in the middle of the frame; this assumption is confirmed by all the videos in the dataset.
 The mask was generated by identifying the most common color in the image's central columns; the color is represented by a range in the Hue component (with respect to HSV color space).
 Building upon this initial step, Michele exploits the Canny edge detector and OpenCV's HoughLinesP function.

Report/section/tableSegmentation.tex

@@ -1,5 +1,4 @@
 \subsection{Table segmentation}
-
 In order to isolate the table with high precision, Michele employed a combination on two different binary masks through a voting system.
 The involved masks are:
 \begin{description}