I am using the Caffe framework for Ubuntu 14.04 and I want to use caffemodel to classify.
At first,I convert the face database into lmdb format and I have trained the caffemodel(lenet_iter_10000.caffemodel).
Next, I try to use my caffemodel to classify.
I read the codefrom google caffe users and rewrite some of the content , but I have some errors.
'AddMatVector' is not a member of 'caffe::MemoryDataLayer'
I asked in some forums, but I did not receive a reply.
Could anyone help me?
If my description is not clear enough, I will make it clearly. Thank you.
I don't see any AddMatVector in the documentation.
But I can see it in the source code in an #ifdef USE_OPENCV. HTH.
Related
For a business process discovery task, I am trying to generate a process model, following pm4py python library. Here's a sample code:
!pip install pm4py
import pm4py
log = pm4py.read_xes('/content/running-example.xes')
process_model, initial_marking, final_marking = pm4py.discover_petri_net_inductive(log)
pm4py.view_petri_net(process_model, initial_marking, final_marking, format="svg")
However, I get output as:
parsing log, completed traces :: 100%
6/6 [00:00<00:00, 121.77it/s]
But no image as is expected from the website: https://pm4py.fit.fraunhofer.de/getting-started-page#discovery
Being relatively new to the world of python, what I learnt from other coders' suggestions here on SO that always read in depth the source code in case of open source libraries.
Here is pm4py visual links:
https://github.com/pm4py/pm4py-core/blob/afee8b0932283b8f8f02dd2b6cc0968a1f1cc723/pm4py/visualization/process_tree/visualizer.py#L69
and specifically for my example:
https://github.com/pm4py/pm4py-core/blob/afee8b0932283b8f8f02dd2b6cc0968a1f1cc723/pm4py/vis.py#L17
But I am not able to figure out how to manipulate it.
Can someone please point out the problem to me and help me generate the views. Also, if anyone has done business process generations before, maybe if you could suggest me any libraries or techniques to analyse event-logs data it would be really helpful.
to visualize the process models mined in PM4Py, make sure that you have graphviz installed on your computer.
see https://pm4py.fit.fraunhofer.de/install for more information on this.
I wanted to know if anyone has managed to use OpenCV in the example codes given by ARToolKit5. I was wishing to use it mainly in the nftSimple example. But since the artoolkit5 code is written in C and using opencv in C is not easy, was wondering if anyone has any ideas, or could give me some sample code to start off with. Thanks in advance!
Check out https://github.com/artoolkitx/artoolkitx/tree/master/Source/ARX/OCVT this is the newer image tracker based on OpenCV and using jpeg image files.
I am completely new to OpenCV, OpenFrameworks, etc. My task is to trace a simple object (a drawing of a circle or a color bottlecap) and display a marker of any kind over the object in the video. I have searched very long and looked through many sites but there is no concise tutorial for a complete beginner, how to build, how to run, what to write and how to configure. Could someone please help me? I'm using Linux Ubuntu.
For setup opencv in eclipse refer this link:
http://docs.opencv.org/2.4/doc/tutorials/introduction/linux_eclipse/linux_eclipse.html.
And for object detection first you need to train data for what you want to detect.For that refer these links,
http://docs.opencv.org/3.1.0/dc/d88/tutorial_traincascade.html#gsc.tab=0
http://docs.opencv.org/2.4/doc/user_guide/ug_traincascade.html
http://docs.opencv.org/2.4/doc/tutorials/objdetect/cascade_classifier/cascade_classifier.html#cascade-classifier
Hope this will help.
I haven't found any server side panorama making from stitching images or a video. I would like an open source alternative, but found any. I just don't want to go trough the hassle of developing all this on my own since but paid software usually are closed source and not very flexible.
I've seen some nifty panorama from video software in the iphone and thought it would be easy to find on *nix systems but with no luck.
Any help will be appreciated. Thanks in advance.
The only option I know is to use panostart (which is part of hugin) from whatever server language you are using.
See here for more information and other command line tools that do parts of the process more specifically.
panostart just works with images, so obviously if you wanted it to work with videos then you would have to process the videos with something like ffmpeg -i z.mov -f image2 export2\%d.png to generate images to pass to panostart.
The other alternative which requires more effort is to write something that uses libpano13 and libffmpeg directly.
You can have a look to VideoStitch, a command line is provided to automate the stitching.
I am working on my web service, and required to persist some image (jpg whatever) and video(wmv) into memmory. Just want to use single_node_cluster to feel voldemort.
Can anybody give me a hint of the configuration and sample code of voldemort? I mean how to configure the value type in stores.xml? protobuf? java-serialization?
Any sample or link would be helpful. Thanks
sample code is here. The same code is also in SAMPLE folder of voldemort installation.
Regarding image store in voldemort, you need to encode the image using base64encoding and then save the text generated in to the value.