deletions | additions       diff --git a/where_R_is_the_maximum__.tex b/where_R_is_the_maximum__.tex  index b8e0a57..ac411a9 100644  --- a/where_R_is_the_maximum__.tex  +++ b/where_R_is_the_maximum__.tex 
...
The Kinect motion sensing involves five level processing. In the study of \citet{brandvik2012investigating}, under the data processing are two different processes, namely the Image processing and the Depth processing. The image processing will process the filtering, colors and scaling. The input data will be translated to gray-scale which prepares the image for skeletal and motion tracking. The filtering process will remove and enhance the data, removing the data refers to noise reduction by means of Gaussian filtering enhancement. In the process of scaling, it allows algorithms to focus on a certain area in a frame or decreasing the resolution for execution. The depth processing handles the scaling on the z-axis that will also make the image three dimensional \cite{Ren_2013}. In the second level processing, the feature extraction will be use to compare each frames of the data to correctly detect an object of interest. For the next level, the skeletal tracking of a human body will be use to represent a number of joints with each one represented by a 3d coordinate \cite{zhang2012microsoft}. Furthermore, in the gesture recognition, it is the process of looking at patterns and distinct features of an object in an effort to recognize whether which object it is and what object to be tracked in the scene. In the last level processing, the template matching will be able to find the object in an image and track it by looking at the similarities of a smaller part of the image to determine whether the gesture is valid or not. But, despite of the fact that Kinect can provide an overall human body tracking, it is still a problem to use Kinect, specifically, in hand gesture recognition \cite{Ren_2013}.  The researchers then proposed to use the Collision detection algorithm to further enhance the performance of the Kinect and projectile model. The collision detection algorithm works by tracking the relation of each pair of solids. The analysis problem of collision detection generally occurs as part of the synthesis problem of collision-free path planning. The path planner may be a person working with a robot simulation system or an algorithm computing over a representation of objects and trajectories \cite{culley1986collision} which is a fundamental need of the study. The collision avoidance refers to the ability of an object to avoid undesired collision. The incorporation of the collision detection algorithm in the Kinect and Projectile model is seen in two major game processes namely, trash segregation and catching the trash using the net and the correctness and validity of the gesture as compared to a provided data set  \cite{de2012integrated}.