File: NEWS

package info (click to toggle)
opencfu 3.9.0-3
  • links: PTS, VCS
  • area: main
  • in suites: sid
  • size: 2,484 kB
  • sloc: cpp: 6,777; xml: 2,442; sh: 1,254; makefile: 210; ruby: 13
file content (60 lines) | stat: -rw-r--r-- 3,019 bytes parent folder | download | duplicates (2)
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
2013-08-27: 3.8.11
    Cleanup makefile and code

2013-08-13: 3.8.10
        Fix eroneous detailled output

2013-08-04: 3.8.9
	BUILD:
		fix some configure issues and warnings.
2013-07-26: 3.8.8
    GUI:
        Typos
        Fix bug when overwritting files with files already presents
        Add transparency layer to colour wheel
        Now uses two different classifiers
2013-07-15: 3.8.7
	GUI:
		Implementation of Masks and ROIs
    PROCESSING:
        FIX the bug that gave wrong result on platorms such as ARM (e.g. raspberry pi).
    COMMAND LINE INTERFACE:
		Results are printed in stdout

2013-06-19: 3.8.6
    GUI:
        Writing of tooltips to help users :D
        A few critical bugs have been found and fixed -> more stability
    CLI:
        add threshold mode/type as an argument

2013-06-13: 3.8.5

    THIS IS A BETA VERSION. DO NOT USE FOR PRODUCTION PURPOSES.
    Many important changes have taken place since version 3.3 (the originally published version).
    The modifications are summarised bellow.

    PROCESSING:
        The feature classifier is not a human defined arbitrary decision tree (as described in the paper) any more.
        Instead, OpenCFU now uses a trained Random Forest[Breiman, 2001] as particle filter (~15 morphological features are used).
        It is therefore more efficient, but also more flexible because it is easy to re-train the classifier with a different sample set.
    CODE:
        Although it is still C++, the code has been rewritten from scratch with a much more object oriented programming approach in mind.
        The code should then be more modular and easier to maintain in the future. The software should also be more stable.
        In addition, OpenCFU now uses autoconf/automake (./configure && make) wich should make the code more portable.
        Finally, OpenCFU is now on a git repository(on sourceforge). This also helps to maintain and reuse the program.

    GRAPHICAL USER INTERFACE:
        The user interface has undergone deep modifications. The main changes are:
            1) A new area dedicated to see the result as a table/list
            2) A image displayer that is scalable and can be progressively zoomed in (using mouse scroll)
            3) The ability to right click on colonies to exclude/include them
            4) The interface tells the user when the processing is taking place (progress bar)
            5) Possibility to drag and drop images to load them
            6) The interface should be more aesthetically pleasing and work small screens

        The use of webcam, masks/ROI and a few other features have not been reimplemented yet. This will come back in the future.

    COMMAND LINE INTERFACE:
        OpenCFU can now be properly used from terminal (see the manual using `opencfu -h') and the program can event be compiled without Glib/Gtkmm support (i.e. without a GUI).
        This is useful to run the program on a workstation or a server and to quickly compare/test the result on image sets.