Tuesday, January 25, 2011

Assignment 3 Part 1 - Setting up the pipe


Social Media Photo Firehose - Combines vendors from Flickr, Panoramio, Picasa, Photobucket, DeviantArt. 

Mashiness Performance - BAD


Pipe returned with only 14 results, with results parsed from photobucket only. http://pipes.yahoo.com/pipes/pipe.info?_id=7e6673ecfa5885bff59e6216a17dfd84

Issues:

1) Flickr has over 16000 photos related to "UNSW," either Yahoo! Pipes will time-out due to the number of data in the feed, or Flickr will stop the requests comming from the API. It may not even be feasible to push 16000 photos into Photosynth even if the pipe can parse 16000 results into a feed. 

2) A complex of operations was set up with multiple sites, particularly linking Google spreadsheets to loop call for page numbers. As a result there are weak links in the Pipe. For example, if Google Spreadsheets fails to parse then the rest of the URL will not work. 


Solutions:

1) reduce "mashiness" of the pipe and cut the pipe into individual vendors. 

2) reduce scope of images, perhaps shrink the data sample to highest rated 3000 images of "unsw"




Wednesday, January 19, 2011

XBox Kinect Hack - 3D Room Scanning and Model Construction

Assignment 3 Proposal

Proposal: To combine social media, photogammatery and spatial information systems to determine "Photogenic Densities"




Technologies Used:  Social Media, Yahoo! Pipes, PhotoSynth+Toolkit and 3dsMax.



Social Media And GIS Mashup - Flickr, Panoramio, YouTube Densities.


Social Media and GIS Mashup - Flickr, Panoramio, YouTube Densities based on "likes", Parramatta and Blacktown.


Yahoo! Pipes Social Media Firehose.


Identification of the most photogenic facade in the search "UNSW Library"



My proposal aims to use Yahoo! Pipes to tap into the wealth of bottom-up information available in social media, visualize them using Photosynth, and overlay them onto a geographic plane or a geographic model.

The information can be used to determine desirable photogenic orientations, and/or popular spaces where people has been in. The key is not to get a overall accurate photosynth from social media, but to determine how one face/element/orientation etc is more detailed and popular than the other, thereby allowing us to see a range of densities and allowing us to make a judgement of photogenic facades and desirable/undesirable spaces.

Challenges:

1) Scale/Selection of Survey area(s) - To make a comparison, ie "UNSW" i will have to filter through more than one keyword and one set of synth, ie all the buildings and spaces in the whole UNSW Campus. First thing would be to identify general areas in the UNSW campus worth doing the query. 2 weeks will not cover every building in the entire campus. If i scale it down to an area or several areas, what justification do i have for that?

2) Variation of Detail - Finding a visual/geometric standard to determine photogenic rating. Amount of point cloud per search/ Amount of polygons per mesh?

3) Integrity and mashiness of the Yahoo! Pipe - Does the Yahoo! Pipe perform as expected? Can many different social media vendors (Flickr, Panoramio, Picasa, Photobucket etc)can be joined into one pipe, or are they all on separate file and "unmashed"


AR Museum Way-finding Model - FINAL

Since the slice modifier fails to export into .ive format, ive taken a more brute approach where i simply scaled the model on the right keyframes. The model still suffered from a few setbacks mainly its refusal to animate or add texture to geometries that have many vertices. Nevertheless this is still an improvement from the last try.

AR Museum Way-finding Model - First Pass

Here is a video of the ar model directly converted into .ive format. As you can see around 70% of the animation has been lost, rendering this device less useful.


Tuesday, January 18, 2011

PROBLEMS! Museum Way-finding Aid

Looks like 70% of all the effects i intend to pull out of max is not going to make it into the model. A lot of explanation is required, so i just made a video of all the problems here.


Proposal: AR Museum Way-Finding Aid




For assignment 2 i chose SANAA's 21st Century Museum at Kanazawa as the architecture in question. This museum has exhibition halls scattered throughout the building, producing many paths, and making way-finding difficult.




4 courtyards are integral to this building which brings light into the center and which also acts as visitor beacons. 4 markers are proposed to correspond to these 4 courtyards, which then highlights the paths through the adjacent exhibits linking through to other courtyards. These courtyards and paths will then be accordingly color coded.

My AR object is to function as a way-finding aid in an environment (the museum) in favor incorporating mixed-reality exhibits. Below is a video clip of the 3dsmax animations i intend to export and place on my markers.


Thursday, January 13, 2011

Layar and Social Media Mashup


1. Experiment in bringing my former Yahoo! Pipes of favourited Flickr images into Josh's RSS Gateway.





Link to my Pipe: http://pipes.yahoo.com/pipes/pipe.info?_id=7e6673ecfa5885bff59e6216a17dfd84

Monday, January 10, 2011

Evil George Builds His Own Prison - Final Result


EVIL GEORGE BUILDS HIS OWN PRISON

This model is abstracted from a photogammatery of a George Bush doll. The distortion is based on the concept of the Klein Bottle and the infinite loop. What was once the hands, feet, head and back of George Bush is now stretched to an external shell, which then folds back onto himself. The evil actions coming out from his hands and his head now comes back, like karma, to build his own prison.
















Evil George Builds His Own Prison - Development

The model deviated from the lattice pattern i was experimenting with for two reasons. One was that the lattice modifier presented the worst case scenario STL Check. Second of all was that it completely knocked off any reading of the folding effect which i was trying to achieve as seen in the two images below.


PROCESS





1. Model was Resized to fit within 100mm cube.


2. Using a series of FFD Modifiers the hand, head, and back was attached to the exterior surface of the model as per the original bottom. Vertex snap and welding secured them to the surface.


3. The model was taken through a Optimise Modifier and then a MeshSmooth modifier to decrease and then smooth out the number of polygons. A series of STL Check modifier was stacked under Edit Mesh Modifiers to delete/edit the error polygons. 

The final cleaned surface was then offset to the thickness using the Shell Modifier.  


 4. A series of spheres was placed around the model. A series of Booleans was done to cut the surface using these spheres.



Final Result!

STL Check

Just made a discovery for fixing errors identified in the STl checker.



Adding a Edit Mesh modifier on top of your STL check modifier will allow you to edit the faces that are generating the errors. I mostly get the double faces error, so i would isolate this in the STL check before apply the edit-mesh modifier. Use a combination of Flip or unify, or just delete and redraw the error faces.

Sunday, January 9, 2011

jackyyuen.com is up!

Hey guys just finished my flash portfolio go check it out. 


Click on the image to be redirected or just go to
http://www.jackyyuen.com

Friday, January 7, 2011

Evil George - Builds his own prison


Evil George Builds his Own Prison

My sculpture aims to draw my George Bush photosynth (George Bush doll from Russell Lowe) and create a prison around him. An interesting thing occurred from photosynth noise where the synthed background attempted to close the model onto itself as shown below. From there i came with the idea of the Klein Bottle and the infinite fold, and using that form to invoke karma between George Bush and his prison (what goes around comes around). 

Source: Uncleaned photosynth.






















Using Modifiers Optimize and Lattice in 3dsMAX. I aim to attach his arm over the lattice in the weekend to achieve the Klien Bottle effect.