Home
Name Modified Size InfoDownloads / Week
Additional 2014-03-14
SOM-Qt-Sample.ZIP 2014-09-10 29.5 MB
readme.txt 2014-09-08 2.5 kB
python-nn.zip 2014-03-19 829.9 kB
Totals: 4 Items   30.3 MB 0
The following is a log of changes in down-up order. That is, the last change description is on top.

-----------------------
8.9.2014
-----------------------
1. New QT sample with SOM
2. Reimplemented Fast DCT
-----------------------
14.3.2014
-----------------------
1. Implement Hopfield network
2. Implement one more test-case:
An attempt to repeat experiment performed by Sherif Saif, Hazem M. Abbas, Salwa M. Nassar and Abdelmonem A. Wahdan, "An FPGA implementation of a neural optimization of block truncation coding for image/video compression", Microprocessors and Microsystems, Volume 31, Issue 8, 3 December 2007, Pages 477-486, Special Issue on FPGA-based Reconfigurable Computing.
3. New script "psnr.py"
It is useful in evaluating image results.
4. Save .lnk files to run Python from the test directory
When debugging Python, we can do "from test import *".
In addition, all relative paths are the same as in script.
-----------------------
9.2.2014
-----------------------
1. Fix SOM2d UpdateWeights loop. Otherwise it is not useful at all.
The same bug is found in BatchSOM2d and in new URGLVQ.
2. Update distribution scripts
3. Implement one more test case:
An attempt to repeat experiment performed by Zhou Shui-shenga, Wang Wei-wei and Zhou Li-hua, "A new technique for generalized learning vector quantization algorithm", IEEE APCCAS 2000 2000 IEEE AsiaPacific Conference on Circuits and Systems Electronic Communication Systems Cat No00EX394, Volume 24, Issue 7, 2006, pages: 339-344.
-----------------------
14.10.2013:
-----------------------
1. Fix non-initialized types.
Otherwise, all classes except SOM couldn't be used from Python.
2. Update args filter
-----------------------
08.10.2013:
-----------------------
Work-around for a bug in PIL:
- Can't save a valid TIFF image in "I" (gray) mode
- Convert into "RGB" first
-----------------------
07.10.2013:
-----------------------
Python package with working test.
This is an attempt to repeat experiment performed by Megha Sharma and Rashmi Kuamri, "A technique of image compression based on discrete wavelet image decomposition and self organizing maps", International Journal of Computer Engineering Science, Volume 3, Issue 8, June 2013.
It's in directory tests/Megha_Rashmi_2013.
The test is based on
    "arcode" - for arithmetic coding
    "PIL" - for generic image processing
-----------------------
New icon
Taken from http://www.kurzweilai.net/images/neuron-640x480.jpg
That is crediting UC Berkeley.
Source: readme.txt, updated 2014-09-08