How to correctly set all the paths and run the code? about objectness HOT 14 OPEN

I second this question. Not to hijack, but I also cannot find any mention of "WinRecall.m", which is used in main.cpp. It seems related to the issue of setting up all paths and downloading the required files.

from objectness.

@sunshineatnoon No, you do not need to download ImageNet unless you want to try BING on ImageNet. Otherwise, stick to VOC.

from objectness.

@phil0stine Sorry for the confusion. The answer is NO. In fact, WinRecall.m is generated from the BING program.

from objectness.

@sunshineatnoon , I have the following error similar to yours, do u know how to solve it??
Invalidate class name
in /media/BING_Linux/Src/DataSetVOC.cpp:125
OpenCV Error: Assertion failed (Invalidate class name
) in loadBox, file /media/BING_Linux/Src/DataSetVOC.cpp, line 125
terminate called after throwing an instance of 'cv::Exception'
what(): /media/BING_Linux/Src/DataSetVOC.cpp:125: error: (-215) Invalidate class name
in function loadBox

from objectness.

@sunshineatnoon @bittnt got the same problem. I'm trying to run for it exam. Any solutions? :\

from objectness.

This means the VOC files you are trying to read do not meet the requirement for the code. Check the DataSetVOC.cpp line 125, probably because you have extra space or invisible symbols in that txt file.

from objectness.

how do u run bing on a given image. not VOC dataset. Forexample: how od you get the 2D object proposals in bing on lena.png

from objectness.

@RyanCV Hi Ryan, did you find how to solve this problem? I have tried opencv2.4.13 and opencv3.2.0, both of them report this message to me. Thank you.

from objectness.

@bittnt @RyanCV @sfandi I encounter the same problem. Has anyone found a solution ?

I print the path to the VOC dataset /mnt1/logs/caffe/objpro/BINGpp/VOCdevkit/VOC2007/ImageSets/Main/train.txt
I also print trainNum and testNum
trainnum: 2501
testnum: 200

Yet the error:
Invalidate class name
in /mnt1/logs/caffe/objpro/Objectness/Src/DataSetVOC.cpp:131
OpenCV Error: Assertion failed (Invalidate class name
) in loadBox, file /mnt1/logs/caffe/objpro/Objectness/Src/DataSetVOC.cpp, line 131
terminate called after throwing an instance of 'cv::Exception'
what(): /mnt1/logs/caffe/objpro/Objectness/Src/DataSetVOC.cpp:131: error: (-215) Invalidate class name in function loadBox

Aborted (core dumped)
Which txt are you talking about @bittnt when referring to extra space and invisible symbols ??

from objectness.

Check the encode type for the txt files.

from objectness.

Hi , Shuai Zheng @bittnt , thanks for your work. I do not understand what you said above "Check the encode type for the txt files.", I think i have a right path for VOC2007 and right format "yml" instead of "xml" , but i also get the same problem :

in /home/dorioliu/Documents/objectdetecion/BING/Objectness-master_linuxversion/Src/DataSetVOC.cpp:125
OpenCV Error: Assertion failed (Invalidate class name
) in loadBox, file /home/dorioliu/Documents/objectdetecion/BING/Objectness-master_linuxversion/Src/DataSetVOC.cpp, line 125
terminate called after throwing an instance of 'cv::Exception'
what(): /home/dorioliu/Documents/objectdetecion/BING/Objectness-master_linuxversion/Src/DataSetVOC.cpp:125: error: (-215) Invalidate class name
in function loadBox

from objectness.

Hi, @RyanCV ,did you solve your problem ?

from objectness.

@dorioliu @ming-c @RyanCV
Did you set up the database correctly? It should look like this: https://mmcheng.net/bingreadme/ , you can also check the paths to the training and test files in the constructor of DataSetVOC. For some reason the function CmFile::loadStrList() did not work for me (I am using OpenSUSE) and I modified the code that the files are red correctly.
And also check if you have a class.txt file next to train and test.txt. My download did not contain that, so I wrote one myself. Now the application is able to produce the training models.

from objectness.

@MaryAnnZ is right!
I just comment the line "line.resize(sz - 1)" in function CmFile::loadStrList().
then it works.

from objectness.