Archives
-
2014
-
November
-
Forward propagation
11/22/14
So the forward propagation is ready and working. Unfortunately, due to memory constraints, I had to change the network structure a bit.
One 60x60 sample is processed by a 2x2 group of processors. Given that Parallella can have 4x4 or 8x8 grids, this allows 4 or 16 groups respectively work -
As a very quick update
11/13/14
Within the last week or so
- downloaded the datasets that were used in 300 Faces in the Wild.
- implemented basic functions like image pre-processing, convolution + maxpooling etc.
Still not sure what to do with unshared convolution, plus there are some other questions that were -
The board
11/02/14
So the board works fine, and apparently worked fine all the way.
Except for the USB bug. Apparently to get USB working I'll need to vandalize a USB cable.
Might try that some time, but for now I've connected it to my home router and SSH'ed into it. Examples run f
-
Forward propagation
-
October
-
Some papers I've been reading
10/30/14
Last week was quite busy with midterms (plus there was nobody in the lab, and I don't have keys :) so I wasn't able to start fiddling with the board. Instead, since there is an obvious winner among the algorithms, I decided to start looking at it.
The basis is described here: ht -
Nicer test result view
10/30/14
The image set can be downloaded from here.
Full test results as Excel file. -
Test results
10/15/14
It has taken more time than I anticipated, partly because I wasn't able to do much during IAC, partly because the existing face detection demos are often hard to find/hard to setup/no longer can be found where they were supposed to be etc. As a result, out of the programs I found, I was only ab -
Final dataset
10/01/14
The final dataset composition:
120 images taken in the lab
90 images of dolls from internet
10 photos from actual rescue missions
80 images of rubble, wood, boxes etc.
-------
300 in total.
Will post the info about the testing when I return from the Int
-
Some papers I've been reading
-
September
-
Forming image set
09/24/14
Images of the victims taken yesterday, total ~140, ranging from easy to trickier.
If I cannot find enough pictures online to increase the set size to at least ~500, I will need to buy a few more dolls, because that is about all non-repeats I can get with these three. -
It works!
09/18/14
Well, kinda. Good news: it boots.
Bad news: usb doesn't work. Todo: double check the drivers. -
Victims
09/17/14
Dolls from Value Village, seemed the most similar to the ones used in the competition. -
Test entry
09/16/14
Journal for COMP 4520 Project.
This is a test entry.
-
Forming image set
-
November