-
Notifications
You must be signed in to change notification settings - Fork 0
/
emlun.log
154 lines (121 loc) · 6.09 KB
/
emlun.log
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
2012-05-10
Worked on making gwhisker tracking work. After a while we had a breakthrough when the tracking started behaving properly. Precision is still quite bad, but it is no longer completely random and actually seems to try to track the gwhisker. The critical change was to increase the exponent of the weight function in the database from 2 to 4 (Euclidean distance^4 instead of ^2).
2012-05-09
(Log written 2012-05-11)
Finished first version of gwhisker generation.
Started working on tracking gwhiskers.
Had a meeting with Örjan, see protocol.
2012-05-08
(Log written 2012-05-09)
Did a little work on the presentation.
Tweaked square tracker, now works reasonably well.
Started working on generation of whisker videos.
2012-05-03
Worked on the mechanical model of whiskers. Tried plotting splines against real whiskers, it seems 4th order polynomials are good enough for representing whiskers.
2012-05-02
(Log written at 00:20, 2012-05-03)
Committed some small things I'd been working on.
Integrated "simple" and "online" square generators into the generator system.
Started working on mechanical model of whiskers using variational calculus.
2012-05-01
(Log written 2012-05-02)
Refactorizations. Added CLI for generators.
2012-04-30
(Log written 2012-05-02)
Refactorizations.
2012-04-29
(Log written 2012-05-02)
Worked on the abstract.
Minor refactorizations.
Made generators capable of generating movies with multiple objects, and made the benchmark track them all in sequence.
2012-04-27
Built generalized cliutils and made TrackerBenchmark use it.
Fixed an error in pf: weights were calculated for X_{t-1} and not for X_t.
Worked a bit on parallellizing tracking.
(The following written 2012-05-02
Added SquareTrackerBetterGoodness which uses Jim's image response thingies
2012-04-24
Built CLI for TrackerBenchmark
2012-04-20
Improved on the benchmarking engine.
Created a general framework for building trackers.
Created generators of test cases for and a tracker for simple pendulums.
2012-04-19
Refactored test generators.
Built a framework for testing trackers and displaying results.
2012-04-18
(Log written 2012-04-19)
Worked more on the square tracker. Decreasing "dt" in the tracked data made it
work a bit better.
Also realized a significant improvement to the database could be done, and
implemented this. The result was that no cartesian products need to be computed
and manipulated, which takes complexity down from exponential in the number of
independent parameter groups to linear.
2012-04-17
Performed our INSWIN presentation.
Implemented some of Jim's work into the square tracker, and worked on trying to get it to work. Not much progress.
2012-04-16
(Log written 2012-04-17)
Prepared our INSWIN presentation.
2012-04-13
Rushed the report for presenting on the workshop. Made quite some progress, actually.
2012-04-12
Worked on the theory part of the report.
Wrote informal text on the particle filter.
Started on illustration for the whisker parameterization.
2012-04-11
(Log written 2012-04-12)
Wrote a little on the report, mostly the theory part.
Worked on making square_tracker_bounce, didn't make much apparent progress.
2012-04-04
Worked more on the database, got a bunch of stuff working.
2012-04-03
Worked on the database refactoring. Made unit tests for wdb.
2012-03-30
(Log written 2012-04-03)
Worked on generation of bouncing squares.
Started working on splitting the database backend into groups of independent parameters.
2012-03-29
(Log written 2012-03-30)
Tweaked bouncing squares generator.
Started working on optimizing database queries and implementing the PF on the bouncing squares.
2012-03-26
Wrote a bit on the report together with Jim.
Generated falling, bouncing squares and ran the DB tracker on them. Horribly bad results and time performance.
2012-03-23
(Log written 2012-03-26)
Improved some docstrings. Fiddled a bit with the square trackers. Talked with Jim a bit about the project in general.
2012-03-09
Continued work on the database:
- Improved generation of database in the square generator
- Added sampling functions to the database
- Implemented a square tracker that uses the database and tried some different parameter values
Eric Larsson stopped by to watch the tracker in action and offer some of his thoughts.
2012-03-01
Refactored the trivial PF implementation to save images to disk rather than show them on screen
Some other refactoring
Worked on state transition database
2012-02-29
Wrote function for creating SQLite database with tables, also worked on abstraction layer but didn't commit anything.
Monday meeting (yes, today is Wednesday): see protocol.
2012-02-23
Christropher informed us that he will switch to another project group. The reason was that the other group had received a project he had asked for.
Finished the trivial particle filter implementation. Can now track a moving, rotating square well enough.
Fixed some errors in wmath.distribution and wmedia.video.
2012-02-20
Monday meeting, see protocol.
Worked on the simple implementation of the particle filter.
2012-02-17
Tweaked plan a bit.
Discussed project plan and implementation details.
Wrote particle filter core functions.
Wrote weighted_choice function and some tests for it.
2012-02-16
Between 10:00 and 12:00: Read more about the particle filter, wrote down the algorithm to better understand it, tried to construct a simple tracking example. Helped Christopher with Ubuntu. Accepted Jim's proposed project plan.
Between 14:00 and 17:00: Finalized first draft of time plan with the other two. Discussed how the particle filter works. Discussed some implementation details.
2012-02-15
Between 10:00 and 12:00: Read about particle filter, instructed Christopher in the use of git, and read some more.
Between 15:00 and 17:00: Inspected a simple third party Python implementation of the particle filter, discussed the project a bit, read a lot more about the particle filter.
2012-02-13
Did complimentary assignment for missing lecture 2012-02-08 for ~4 hours.
Monday Meeting with Jim and Christopher for 2 hours (they started some 1.5 hours earlier). Goal of the week is to complete a particle filter implementation and some fabricated input data.