Computer Vision Coding Example: Java Multicores

CMSR ML Studio generated program source codes are very easy to integrate into your applocations. This page will show how to code to use Computer Vision functions. Depending on modeling types, calling parameters are different.

Model Code Generation and Compilation

From CMSR ML Studio, perform the followings to generate model codes;

1. Open your computer vision model.
2. From "Modeling" menu, select "Export model as program code".
3. Select "Language" as "CPU: Java Multicore" and press the "Copy" button. This will generate Java program file to your system clipboard. Create a model file into your project folder and copy clipboard content and save, say, "CMSRModel.java". Note that generated program will access base class files. This class definitions will be stored at "rosellacnngpu.jar" file. Add it to your class path. The jar file can be found from the "CMSR_HOME/cmsrsystem/resources" folder.

Copy a model parameter file into locations where your main program can access. And modify access path in your main program. To create model parameter file, from "Modeling" menu, select "Export model as file (MyDataSay, BI Server, CNN, ...etc.)".

CNN/FCN Classification Example

The following code shows how CNN/FCN classification is called. You can control the number of CPU threads to use. Normally it is equal to the number of CPU cores or CPU hyper threads. Note that CMSRModel is the model class that CMSR Studio generates.

import java.io.File;
import java.io.FileInputStream;

public class Eval1 {

public static void main(String[] args) {
		
	String filebase = "C:\\YourPath...\\";
	
	int numberOfThreads = 4;

	String modeltype, modelfilepath;
	int[][][] image;
		
	boolean blackandwhite = false;
	boolean r0g1b2 = true; 

	CMSRModel model = new CMSRModel();
		
	try {
		modeltype = "CNN";
		modelfilepath = filebase+"\\modelfile."+modeltype.toLowerCase();

		// initalize resource
		model.initializeModel(numberOfThreads, 
				new FileInputStream(modelfilepath)); // must be an InputStream
		
		// repeat the followins as needed. can get image from camera.
		File testimagefile = new File(filebase+modeltype+"images.jpg");
		image = getRGBArray(testimagefile, model.input_image_width, model.input_image_height);

		int outLabelCount = 4;
		int[] outLabelIndices = new int[outLabelCount+1]; // notice extra 1
		float[] outLabelProbabilities = new float[outLabelCount+1]; // notice extra 1
			
		model.evaluate(outLabelCount, outLabelIndices, outLabelProbabilities, blackandwhite, r0g1b2, image);
		for (int i=0; i < outLabelCount; i++) {
			System.out.println(""+i+": "+outLabelIndices[i]+"="+outLabelProbabilities[i]);
		}

		// release resources;
		model.assignNulls();

		System.out.println("Done.");
	} catch (Throwable e1) {
		e1.printStackTrace();
	}
}
}

M-CNN Multi-value Output Regression Example

The following code shows how M-CNN multivalue output regression is called. You can control the number of CPU threads to use. Normally it is equal to the number of CPU cores or CPU hyper threads. Note that CMSRModel is the model class that CMSR Studio generates.

import java.io.File;
import java.io.FileInputStream;

public class Eval1 {

public static void main(String[] args) {
		
	String filebase = "C:\\YourPath...\\";

	int numberOfThreads = 4;
		
	String modeltype, modelfilepath;
	int[][][] image;
		
	boolean blackandwhite = false;
	boolean r0g1b2 = true; 

	CMSRModel model = new CMSRModel();
		
	try {
		modeltype = "MCNN";
		modelfilepath = filebase+"\\modelfile."+modeltype.toLowerCase();

		// initalize resources;
		model.initializeModel(numberOfThreads, 
				new FileInputStream(modelfilepath)); // must be an INputStream

		// repeat the followings as needed. can get image from camera;
		File testimagefile = new File(filebase+modeltype+"images.jpg");
		image = getRGBArray(testimagefile, model.input_image_width, model.input_image_height);

		int outLabelCount = model.getClassCount();
		float[] outLabelProbabilities = new float[outLabelCount];
			
		model.evaluate(outLabelProbabilities, blackandwhite, r0g1b2, image);
		for (int i=0; i < outLabelCount; i++) {
			System.out.println(""+i+": "+outLabelProbabilities[i]);
		}

		// release resources;
		model.assignNulls();

		System.out.println("Done.");
	} catch (Throwable e1) {
		e1.printStackTrace();
	}
}
}

OD-CNN Object Detection Example

The following code shows how OD-CNN object detection is called. You can control the number of CPU threads to use. Normally it is equal to the number of CPU cores or CPU hyper threads. Note that CMSRModel is the model class that CMSR Studio generates. The X/Y coordinates are the center locations of detected object bounding boxes. Output width and height are relative size where whole size is 1.0.

import java.io.File;
import java.io.FileInputStream;

public class Eval1 {

public static void main(String[] args) {
		
	String filebase = "C:\\YourPath...\\";

	int numberOfThreads = 4;
		
	String modeltype, modelfilepath;
	int[][][] image;
		
	boolean blackandwhite = false;
	boolean r0g1b2 = true; 

	CMSRModel model = new CMSRModel();
		
	try {
		modeltype = "ODCNN";
		modelfilepath = filebase+"\\modelfile."+modeltype.toLowerCase();

		// initialize resources;
		model.initializeModel(numberOfThreads, 
			new FileInputStream(modelfilepath)); // must be an InputStream

		// repeat the followings as needed. can get image from camera.
		File testimagefile = new File(filebase+modeltype+"images.jpg");
		image = getRGBArray(testimagefile, model.input_image_width, model.input_image_height);

		int  maxOutCount = 10;
		int[]   outClassIndex = new int[maxOutCount+1]; // notice extra 1
		float[] outClassProbability = new float[maxOutCount+1]; // notice extra 1
		float[] outX = new float[maxOutCount+1]; // notice extra 1
		float[] outY = new float[maxOutCount+1]; // notice extra 1
		float[] outWidth = new float[maxOutCount+1]; // notice extra 1
		float[] outHeight = new float[maxOutCount+1]; // notice extra 1
			
		int outcount = model.evaluate(maxOutCount, outClassIndex, outClassProbability, 
				outX, outY, outWidth, outHeight,
				blackandwhite, r0g1b2, image);

		for (int i=0; i < outcount; i++) {
			System.out.println(""+i+": "+
				outClassIndex[i]+"="+outClassProbability[i]+"/"+
				outX[i]+"="+outY[i]+"/"+
				outWidth[i]+"="+outHeight[i]
				);
		}

		// release resources;
		model.assignNulls();

		System.out.println("Done.");
	} catch (Throwable e1) {
		e1.printStackTrace();
	}
}
}

T-CNN Similarity Regression/Face Recognition Example

The following code shows how T-CNN similarity regression is called. You can control the number of CPU threads to use. Normally it is equal to the number of CPU cores or CPU hyper threads. Note that CMSRModel is the model class that CMSR Studio generates.

import java.io.File;
import java.io.FileInputStream;

public class Eval1 {

public static void main(String[] args) {
		
	String filebase = "C:\\YourPath...\\";

	int numberOfThreads = 4;
		
	String modeltype, modelfilepath;
	int[][][] image;
		
	boolean blackandwhite = false;
	boolean r0g1b2 = true; 

	CMSRModel model = new CMSRModel();
		
	try {
		modeltype = "TCNN";
		modelfilepath = filebase+"\\modelfile."+modeltype.toLowerCase();

		// initialize resources;
		model.initializeModel(numberOfThreads, 
			new FileInputStream(modelfilepath)); // must be an InputStream

		// repeat the followings as needed. can get images from camera.
		File testimagefile1 = new File(filebase+modeltype+"images1.jpg");
		File testimagefile2 = new File(filebase+modeltype+"images2.jpg");
		image = getRGBArray(testimagefile1, model.input_image_width, model.input_image_height);
		int[][][] image2 = getRGBArray(testimagefile2, model.input_image_width, model.input_image_height);

		float f = model.evaluate(blackandwhite, r0g1b2, image, image2);
		System.out.println("value="+f);

		// release resources;
		model.assignNulls();

		System.out.println("Done.");
	} catch (Throwable e1) {
		e1.printStackTrace();
	}
}
}