.tags

Locomotion	/Users/alvaromartinezfernandez/Documents/UiT/Masteroppgave/thesis/chapters/bignetvr.tex	/^\\end{figure}%$/;"	subsubsection	line:48
Scalable network in Unity and data structures	/Users/alvaromartinezfernandez/Documents/UiT/Masteroppgave/thesis/chapters/bignetvr.tex	/^Finally, our dataset has information about the relationships between the nodes. BigNet VR is implemented to show also this information. Because there can be too many relationships in the dataset, we don't show them all at the same time. Therefore we can only see those of the node that the user has selected. The way that these relationships are represented is with lines between the nodes.$/;"	section	line:100
Other features of BigNet VR	/Users/alvaromartinezfernandez/Documents/UiT/Masteroppgave/thesis/chapters/bignetvr.tex	/^The algorithm used to cluster the nodes in the network is another important aspect that can influence in the scalability. In BigNet VR we use a linear algorithm that clusters the nodes in the 3d space depending on the module where they belong too. In this way the user can visualize each module as single clusters with a distinct colour per cluster.$/;"	section	line:165
New requirements based on the interviews	/Users/alvaromartinezfernandez/Documents/UiT/Masteroppgave/thesis/chapters/futureWork.tex	/^Finally, we would also like to run other experiments to further evaluate the performance of the application and study better the scalability. We would like to use larger datasets in order to determine where is the size limit for which we can visualize these datasets in GeneNet VR for both PC and the Oculus Quest headset.$/;"	section	line:15
MIxT in VR	/Users/alvaromartinezfernandez/Documents/UiT/Masteroppgave/thesis/chapters/mixt.tex	/^\\end{figure}$/;"	section	line:30
Network characteristics	/Users/alvaromartinezfernandez/Documents/UiT/Masteroppgave/thesis/chapters/mixt.tex	/^We have added the possibility to compare two datasets in real-time in GeneNet VR. This can be very useful for bioinformaticians when they work with several datasets like in MIxT. To compare two networks, the user can use a UI slider to visualize two datasets at the same time, creating what we call a morphing effect. To help the user distinguish from one dataset to another, we use linear interpolation for both the color of the nodes and their position.$/;"	section	line:39
Virtual Reality Chemical Space	/Users/alvaromartinezfernandez/Documents/UiT/Masteroppgave/thesis/chapters/related-work.tex	/^I will focus in this chapter on similar works found in the literature for the visualization of bioinformatics data in VR.$/;"	section	line:1
csv	/Users/alvaromartinezfernandez/Documents/UiT/Masteroppgave/thesis/scripts/cdf.py	/^import csv$/;"	namespace	line:1
collections	/Users/alvaromartinezfernandez/Documents/UiT/Masteroppgave/thesis/scripts/cdf.py	/^import collections$/;"	namespace	line:2
number_nodes	/Users/alvaromartinezfernandez/Documents/UiT/Masteroppgave/thesis/scripts/cdf.py	/^number_nodes = list()$/;"	variable	line:4
number_edges	/Users/alvaromartinezfernandez/Documents/UiT/Masteroppgave/thesis/scripts/cdf.py	/^number_edges = list()$/;"	variable	line:5
cum_edges	/Users/alvaromartinezfernandez/Documents/UiT/Masteroppgave/thesis/scripts/cdf.py	/^cum_edges = list()$/;"	variable	line:6
total_num_edges	/Users/alvaromartinezfernandez/Documents/UiT/Masteroppgave/thesis/scripts/cdf.py	/^total_num_edges = 0$/;"	variable	line:7
previous_value	/Users/alvaromartinezfernandez/Documents/UiT/Masteroppgave/thesis/scripts/cdf.py	/^previous_value = 0$/;"	variable	line:8
csv_reader	/Users/alvaromartinezfernandez/Documents/UiT/Masteroppgave/thesis/scripts/cdf.py	/^    csv_reader = csv.reader(csv_file, delimiter=' ')$/;"	variable	line:11
f	/Users/alvaromartinezfernandez/Documents/UiT/Masteroppgave/thesis/scripts/cdf.py	/^f = open(".\/data\/bloodEdgesCDF.dat", "w+")$/;"	variable	line:24
Challenges and research problem	/Users/alvaromartinezfernandez/Documents/UiT/Masteroppgave/thesis/chapters/introduction.tex	/^\\textbf{Thesis statement: } \\emph{Virtual Reality is advantageous for the visualization of large biological networks and for rapid exploration of patterns in them using affordable hardware}.$/;"	section	line:36
Proposed solution and contribution	/Users/alvaromartinezfernandez/Documents/UiT/Masteroppgave/thesis/chapters/introduction.tex	/^The interactivity in VR is usually done with the VR controllers, which simulate our hands in the virtual world. We can implement natural actions for the operator such as grabbing objects or items with their hands, which will feel intuitive. However, we can also implement other actions in the virtual world like using laser pointers to select something, using 2D menus inside the virtual world or teleporting the user to other parts of the virtual space. In GeneNet VR, we are dealing with abstract information and the amount of data can escalate quickly if we do not implement practical solutions. We must therefore maintain a proper balance between the amount of data being visualized, comfort and user-friendly interaction solutions and good performance.$/;"	section	line:71
Outline	/Users/alvaromartinezfernandez/Documents/UiT/Masteroppgave/thesis/chapters/introduction.tex	/^\\end{figure}$/;"	section	line:92
Interaction with the network	/Users/alvaromartinezfernandez/Documents/UiT/Masteroppgave/thesis/chapters/genenetvr.tex	/^Finally, for the \\textit{Teleportation and snap rotation component} we have used some scripts and prefabs that come with the VRTK library. I used as a reference some techniques to develop these locomotion solutions from a course on Unity\\footnote{https:\/\/learn.unity.com\/course\/oculus-vr}.$/;"	section	line:96
Locomotion	/Users/alvaromartinezfernandez/Documents/UiT/Masteroppgave/thesis/chapters/genenetvr.tex	/^\\end{figure}%$/;"	subsubsection	line:114
Scalable network in Unity and data structures	/Users/alvaromartinezfernandez/Documents/UiT/Masteroppgave/thesis/chapters/genenetvr.tex	/^Finally, our dataset has information about the relationships between the nodes. GeneNet VR is implemented to show also this information. Because there can be too many relationships in the dataset, we don't show them all at the same time. Therefore, we can only see those of the node that the user has selected. The way that these relationships are represented is with lines between the nodes.$/;"	section	line:166
Other features of GeneNet VR	/Users/alvaromartinezfernandez/Documents/UiT/Masteroppgave/thesis/chapters/genenetvr.tex	/^The algorithm used to cluster the nodes in the network is another important aspect that can influence in the scalability. In GeneNet VR we use a linear algorithm that clusters the nodes in the 3d space depending on the module where they belong too. In this way the user can visualize each module as single clusters with a distinct colour per cluster.$/;"	section	line:231
Implementation details	/Users/alvaromartinezfernandez/Documents/UiT/Masteroppgave/thesis/chapters/genenetvr.tex	/^\\end{figure}$/;"	section	line:287
Methodology of experiment setup	/Users/alvaromartinezfernandez/Documents/UiT/Masteroppgave/thesis/chapters/evaluation.tex	/^\\end{enumerate}$/;"	section	line:30
Translating the network meets the 72 FPS	/Users/alvaromartinezfernandez/Documents/UiT/Masteroppgave/thesis/chapters/evaluation.tex	/^The average of all the frames is the most important value to look at because it can give us an idea of whether the experiment meets the required FPS or not. In a perfect application where all the frames last the same amount of time and where the frame rate is 72, each frame would last around 13.9 milliseconds. However, this is not the case in real applications. If the general average is above 13.9 milliseconds, it means that the experiment didn't pass the 72 FPS. By looking at the 0.25\\% worst frame time and the 1\\% worst frame time, we will get the frame with the worst time and the 7 frames with the worst times respectively (our experiments last for 700 frames). If we see that these frames have much worse times, compared with the general average, we will use a profiler to see if there are bottlenecks and try to give a solution to them.$/;"	subsubsection	line:122
Selecting nodes the network meets the 72 FPS but with possible bottlenecks	/Users/alvaromartinezfernandez/Documents/UiT/Masteroppgave/thesis/chapters/evaluation.tex	/^The average time for this experiment is also under the 13.9 milliseconds limit, meaning that the interaction reaches 72 FPS. The low 1\\% is also under 13.9 milliseconds, so we don't consider that there are any big issues for the performance of this interaction. The variation among the network sizes is also insignificant. An evaluation with bigger networks would be necessary to determine the network size limit for which the performance is not good.$/;"	subsubsection	line:165
How do users perceive the visualization of large biological networks in GeneNet VR	/Users/alvaromartinezfernandez/Documents/UiT/Masteroppgave/thesis/chapters/evaluation.tex	/^From the results, we can say that we reach 72 FPS using the Oculus Quest for our network size. However, the performance is on the limit of 72 FPS and when we select nodes with more than 10 edges, the performance can get worse. It would be necessary to evaluate GeneNet VR using larger networks on the Oculus Quest. Also improve the creation of the edges as a better programming solution, as we mentioned before.$/;"	section	line:366
How do you perceive the application	/Users/alvaromartinezfernandez/Documents/UiT/Masteroppgave/thesis/chapters/evaluation.tex	/^\\end{table}$/;"	subsubsection	line:397
What is missing in the application	/Users/alvaromartinezfernandez/Documents/UiT/Masteroppgave/thesis/chapters/evaluation.tex	/^We also received some feedback about how to improve the visualization for pattern finding. For example, it would be useful to show the names of the connected nodes to the node that we are selecting. Also, the possibility to isolate or highlight the nodes that we are selecting would be helpful. Finally, pattern finding would also be easier if they had the chance to change the layout of the network.$/;"	subsubsection	line:411