Yesterday: Constructed physical simulation of balls bouncing on a table, and used application to see about object segmentation. Sometimes works. (video, fast) Many times has erroneous identification. Occassionally marks all vertices in a single group, while other times it marks all vertices in separate groups (which is definitely wrong, given a table should be static). Need to check for bugs:

• Are the vertices getting reliably re-ordered?
• What do the block adjacencies look like when finding all vertices the same? When finding all vertices different?
• Are there other bugs in the program?

Provided example for 2, 10 and 49 balls dropping. Worsens as more balls are introduced. Tweaked application so that only current files needed are loaded into memory.

Today: Look into bugs from yesterday. Chat with supervisors about how it looks. Create series of synthetic double pendulum example. Implement transformation-based validation (ie. modification of original step 2 of process), and process for reliance on prior frames. Generate a file-loading for test cases so there is consistency in process validation.

Write up (mathematical foundation, computational/theoretical implementations) for future writing/papers:

• distance-based segmentation
• block-based segmentation/checking

Consider laying out how I'd like the thesis to look and how I'd like it to describe the various implementations and stages along the way. Read more...

Yesterday: Reviewed where I'm at in project development. Built a synthetic example of two boxes rotating about one another. They appear to successfully find each other. (video)

Re-scheduled UX validation test for a colleague. Spent much of the afternoon assisting a colleague with some of their academic challenges.

Today: Create series of synthetic examples from blender and possible double pendulum. Implement transformation-based validation (ie. modification of original step 2 of process), and process for reliance on prior frames. Generate a file-loading for test cases so there is consistency in process validation.

Write up (mathematical foundation, computational/theoretical implementations) for future writing/papers:

• distance-based segmentation
• block-based segmentation/checking

Consider laying out how I'd like the thesis to look and how I'd like it to describe the various implementations and stages along the way. Read more...

Yesterday: (Friday, Dec 12) Motivated work by tweaking Gallery on website, looking into HTML5, javascripting, jQuery, image carousels to consider developing an example countdown timer for desired dates/times. Details on the given page.

This helped motivate me to tackle the block-based separation (see image), including visualizing resulting adjacency. Assumes there are at least 4 solid points in an existing object. Only relies on prior frame to-date. Block checking involves O(nk) space and time, which can be reduced by more intelligent checking.

(Over break) Found a simple blender doll to help with building synthetic examples over time. Modified code base to iteratively load in different obj files, where vertices are in-order and corresponding. This is akin to incorporating animation files, where points are transformed across keyframes over time.

Incorporated logic for segmenting based on block adjacency. ie. Determining which vertices belong to which core block. Added support for showing edges from block adjacency segmentation/re-grouping. Need to handle outlier vertices.

Dropped into career counseling at Waterloo. Recognized that there are many options always open, and its a matter of selecting one and pursuing it. One exercise parsed down a long list of values and needs within a job. Alongside the standards needs of a supportive supervisors, clear directions, good salary, job security and leisure time/work-life balance, my top priorities include feeling productive, having a routine, belonging and interacting with people, in an honest cooperative environment. Problem solving comes naturally within computer science, and location is somewhat important - either near mountains, or family and friends - somewhere that I can be involved in and feel a part of a community.

Today: Review the status of the implementation. Re-schedule user experience validation test for a colleague. Create series of synthetic examples from blender manikin and possible double pendulum. Implement transformation-based validation (ie. modification of original step 2 of process), and process for reliance on prior frames. Generate a file-loading for test cases so there is consistency in process validation.

Write up (mathematical foundation, computational/theoretical implementations) for future writing/papers:

• distance-based segmentation
• block-based segmentation/checking

Consider laying out how I'd like the thesis to look and how I'd like it to describe the various implementations and stages along the way. Read more...

Yesterday: Attended successful PhD defense of colleague. Recognized importance of well-established work. Interesting that examiners mostly discussed where the applicant thought future directions might proceed, as opposed to validating the premise behind the presented work.

Talked with career counseling about organizing materials for possible applications. Not exactly as detail-helpful as I would've hoped, as no suggestions on how to arrange graduate research to explore desirable experiences through research process. Will ask at Waterloo, and continue as planned (ie. these daily scrums; exploring and practising technical side of things, etc). Considering introducing the scrum concept to colleagues in the lab to facilitate improved communications.

Today: Implement grouping/clustering based on block separation. Implement transformation-based validation (ie. modification of original step 2 of process). Generate a file-loading for test cases so there is consistency in process validation. Consider implementing/incorporating animation files from CPSC453 where points are transformed across keyframes.

Consider laying out how I'd like the thesis to look and how I'd like it to describe the various implementations and stages along the way. Document last Wednesday's discussion in latex for possible future writing/papers. Read more...

Yesterday: Scheduled career counseling sessions to assist with direction, both at the University of Calgary - this week, and the University of Waterloo next week.

Added widget and implementation to toggle drawing adjacency edges for the 3D points. Added support for multiple objects.

TA-in-residence has been assigned elsewhere, so it is anticipated that a traditional TA position will be assigned. Outside of benefiting and motivating the students (including introducing them to good coding practises), this benefits me by added exposure to various coding decisions.

Today: Attend PhD defense of colleague. Drop in on career counseling session to solicit suggestions for directions.

Implement grouping/clustering based on block separation. Implement transformation-based validation (ie. modification of original step 2 of process). Generate a file-loading for test cases so there is consistency in process validation. Consider implementing/incorporating animation files from CPSC453 where points are transformed across keyframes.

Consider laying out how I'd like the thesis to look and how I'd like it to describe the various implementations and stages along the way. Document last Wednesday's discussion in latex for possible future writing/papers. Read more...

Yesterday: (Tuesday) Wrap up thesis review. Provided decent feedback at the defense dry-run of colleague. Need to ensure better preparation for clear questions next time.

Cleaned up some discussion about applications as an automated patrol mechanism, adding it to the core motivation notes.

Attempted to get Matlab working with the Kinect. Depth is fine, but the colour causes matlab to crash. Bug report sent.

Stumbled across/remembered auto point tracking tool available from Tyson Hedrick at University of North Carolina. May need to skim through the implementation to see how its related to my own objectives (at least for when I'm looking for feature points down the road).

(Wednesday) Met with Faramarz to briefly discuss TA-ship. Forwarded possible internship location for summer 2015.

(Thursday) Mentally trying to figure out how to select blocks appropriately.

(Tuesday) Applied for TA-ship. Emailed to suggest possibly acting as TA-in-residence, though questioning whether I should have something more technical. Group meeting on mathematical explorations in PuPs from group member.

Today: Document approach for segmentation in latex for possible future writing/papers. Visualize graph representation of the point clouds within this process. Tweak/implement transformation-based validation (ie. step 2 of process). Generate a file-loading for test cases so there is consistency in process validation.Read more...

Yesterday: Wrap up thesis review. Provided decent feedback at the defense dry-run of colleague. Need to ensure better preparation for clear questions next time.

Cleaned up some discussion about applications as an automated patrol mechanism, adding it to the core motivation notes.

Attempted to get Matlab working with the Kinect. Depth is fine, but the colour causes matlab to crash. Bug report sent.

Stumbled across/remembered auto point tracking tool available from Tyson Hedrick at University of North Carolina. May need to skim through the implementation to see how its related to my own objectives (at least for when I'm looking for feature points down the road).

Today: Document last Wednesday's discussion in latex for possible future writing/papers. Visualize graph representation of the point clouds within this process. Tweak/implement transformation-based validation (ie. step 2 of process). Generate a file-loading for test cases so there is consistency in process validation.

Meeting with Faramarz. Decide whether to apply for TA-ship next January or not. Read more...

Yesterday: Met with Faramarz and a Mitacs representative to discuss how funding options work within the organization. Reminder email to Kathreen about biological motivation/applications. Emailed researcher at U of A about possible discussions and looked at a few other options (such as the Road Ecology group in Montana, ). Emailed researcher at U of A about thesis access. Stumbled across the ScareCrow Sprinkler from Contech to deter animals from gardens.

Looked again at ISIS quadcopter and currently available 3D capture cameras for custom rig. Contacted ISIS about ballpark costs.

Started developing a list of possible funding options, skillsets that I need to develop and how to go about doing that in a financially viable way (eg. contract/internship at a related company).

Sorted out license keys and re-installed Matlab with the fancy toolboxes. (And then re-installed after getting the wrong version of 2014a rather than 2014b). Tried out Matlab kinect and image processing toolkit.

Today: Wrap up thesis review. Group meeting for defense dry-run of colleague.

Document last Wednesday's discussion in latex for possible future writing/papers. Visualize graph representation of the point clouds within this process. Tweak/implement transformation-based validation (ie. step 2 of process). Generate a file-loading for test cases so there is consistency in process validation.

Apply new toolkits to current work on dynamic 3D (ie. rigid motion segmentation, precursor to articulated motion identification).

Build a list of what I need (resources: financial, technical, human), and how to get them, including how to motivate others.

Decide whether to apply for TA-ship next January or not.

Possibly communicate with:

• ranchers in Alberta to see what the issue is for livestock loss due to wildlife,
• conservation officers in Alberta to see what directions they're taking and if there would be an opportunity there for assisting,
• CP rail and what their financial obligations are for wildlife and livestock re-imbursement.

Read more...

Yesterday: (Wednesday) Really productive meeting with Faramarz about how I approach the distance-based pruning. Brainstormed how to perform the transformation-based pruning, which may no longer be as necessary. Current computation time is O(n) + O( (n²) / k) for k processors.

Emailed Kathreen, ccing supervisors, about possible motivation, the core idea of which is to "Figure out how to pepper spray grizzly bears using automated UAV patrols." (ie. Automated Park Wardens). It seems a little over-the-top, and requiring a lot more expertise than I have, but my chunk relates to building the tools for modeling 3D dynamic animals.

Connected with Matlab people about access to the Image Acquisition Toolbox and the Computer Vision Toolbox.

(Thursday, Friday) Started reading through the thesis of a colleague to provide sound technical questions and feedback.

Today: Continue reading thesis.

Document last Wednesday's discussion in latex for possible future writing/papers.

Install Matlab toolboxes and try them out. Visualize graph representation of these point clouds. Tweak/implement transformation-based validation (ie. step 2 of process). Generate a file-loading for test cases so there is consistency in process validation.

Explore how to properly tackle this project. Consider taking a year's leave to re-group/re-evaluate, and plan out future directions. eg. Work at somewhere I know I can do the work, and treat this as a hobby/side project, or alternatively work at a company or research team that specializes in observation-based 3D modeling working with 3D scanners.

Build a list of what I need (resources: financial, technical, human), and how to get them, including how to motivate others.

Communicate with:

• ranchers in Alberta to see what the issue is for livestock loss due to wildlife,
• conservation officers in Alberta to see what directions they're taking and if there would be an opportunity there for assisting,
• CP rail and what their financial obligations are for wildlife and livestock re-imbursement.

Read more...

Yesterday: Brainstormed how to perform the transformation-based pruning, with no real progress on best ways to tackle it without standard approaches for triplets (or four-point) finding. Started implementation for visualizing adjacencies - ie. the graph - for the points.

Today: Meet with Faramarz to discuss how the Matlab version of distance-based pruning has been implemented, and brainstorm with him on transformation-based. Visualize graph representation of these point clouds. Implement transformation-based pruning. Generate a file-loading for test cases so there is consistency in process validation.

Looking into getting the Kinect running within Matlab (link) and get back to Matlab group about the various toolboxes. Read more...