i have configured opencv 2.4.8 for android with my eclipse . its working fine i have implemented a test example to check it opens opencv loader using folowing code
OpenCVLoader.initAsync(OpenCVLoader.OPENCV_VERSION_2_4_3, this, mLoaderCallback);
then i load a image and converts it to gray scale it successfully converts the image.
Now i want to use more functions of opencv on image like image filtering etc. for kindly suggest me to start doing that and how to use C++ functions in java eclipse android. kindly suggest me any tutorial for that or any example so that i can proceed
You could try the OpenCV examples for android
http://opencv.org/platforms/android/opencv4android-samples.html
Or a basic test from the official website
http://docs.opencv.org/doc/tutorials/introduction/android_binary_package/dev_with_OCV_on_Android.html#dev-with-ocv-on-android
But I used the examples(in the samples folder in the sdk) instead when I was learning how to use OpenCv in Android
One option that you have is to check the C++ Tutorials, and then check this repository where I did some of those tutorials in Android Java.
More tutorials will be added, if you want a specific one let me know.
Related
As the title says im attempting to build TensorFlow Lite for Java Desktop without Android. There doesn't seem to be any official guide or build on the Tensorflow Github / Website. Does anyone know of an existing 3rd party maven/gradle link that achieves this, or a concrete way to build Tensorflow for Java desktop without Android.
I've tried the following link but when I preform the steps, I get various Bazel build errors.
https://gist.github.com/JnCrMx/144f3169bde1da6856815b3419122ab0
I have written a library for Object Detection in Java. You can have a look if you want: JavaTF
If you want you can just follow the instructions I provided on the Readme file, and test the functionality. Currently it does not support TFLite, but the saved_model format is quite good too. You can have a look how things are done, I think for a TFLite model it is not much different.
Let me know if it helps :)
I'm a little stuck here. I'd like to use the JavaCV library to capture a frame from a camera and store the pixels in a matrix. Unfortunately, all the tutorials that I can find only demonstrate how to do this on an ordinary computer, and since this is for an android application I can't use the necessary AWT library that the tutorials point to. Could somebody please either point me to a tutorial or explain a way that allows me to accomplish this on Android?
I'm using Android Studio.
I am using openCV4Android and I know there are lots of questions related with this topic. Most of them (the ones with answer), suggest a solution with C++ and add it to the project in the folder jni + Android.mk, etc.
Can I do something to avoid C++ and call every method from Java?
If it is not possible with OpenCV, I am open to try alternatives.
Thanks.
There is a JavaCV - Java interface to the OpenCV library, it is popular for the Android application.
You will need to use JNI NDK for stitching if you want to do stitching with opencv
I am trying to develop a Android OCR app (for home use).
I had the idea to use the Asprise OCR in the app but instead of the windows version I was gonna use the Linux version.
The problem is, I don't know how to handle the native libraries. They have the extension .SH
I'm developing in Eclipse.
Google recently released an OCR API, so you probably want to use that:
https://developers.google.com/vision/text-overview
Just add the following line to your dependecies:
compile 'com.google.android.gms:play-services-vision:9.2.0'
I would recommend you to use Android Studio instead of Eclipse and related to the OCR library, you could use Tesseract, which is an open source library developed by Google.
Here you have more information and a practical example to follow:
https://solidgeargroup.com/ocr-on-android
That reference may help you:
http://gaut.am/making-an-ocr-android-app-using-tesseract/
Hope this helps
Update 13/04/17: Removed the broken link, will put it back if it goes up again.
I have Java code for face recognition which works fine. There is a problem however if I port the code into an Android application.
I have included the necessary jama and jmf which is required by the Java code into the Android app. The Android code compiles fine, but when I run the app it's force closed!
What might I be doing wrong here? What are the things that I need to keep in mind when I do the porting?
By the way there is no GUI for the Java code.
Porting your java code to Android will usually not work, you need to foloow the guidelines for creating an Android application.
You should start at the Android documentation for help.
http://developer.android.com/guide/topics/fundamentals.html
I'd recommend you take a look at the Android OpenCV project. In their CVCamera sample, they have a nice setup for doing real-time image processing for feature detection etc. I've previously modified it to support face detection using the OpenCV library.
If you want to use your own code instead, I suggest you strip out all the JNI/NDK interop code and substitute your own. At any rate, it's a nice starting point.