shivajivarma / bliffoscope

Bliffoscope

Application which scans for target images against Test Data.

Read full problem statement over here -> Bliffoscope Problem.pdf

Prerequisites

1. Java Runtime (JDK for Development)
2. Gradle Build Tool (https://gradle.org/install/#with-a-package-manager) a. Extract the Gradle zip file. b. Set the environment path variable pointing to ‘bin’ folder of Gradle.

How to run application

1. Go to project folder and run below command.

How to make build

1. Run gradle build command in project folder.

Notes

1. On start of application, test data(TestData.blf) is analyzed against slime torpedo(SlimeTorpedo.blf) and Starship (Starship.blf), with threshold 70 as default and Result is presented in console. (file are present in resource folder)
2. Threshold : Minimum percentage match required to consider a subset of test data as target.
3. Coordinates: target matrix first cell position.
4. Web interface to test the application is available at http://127.0.0.1:8090/

Implementation

1. findTargets method finds all targets in given test data image(Boolean matrix), It basically computes score for each sub matrix of with height and length that of targetImage.

public List<Target> findTargets(Image sourceImage, Image targetImage, int threshold) {
    List<Target> targets;

    if (sourceImage == null) {
        throw new NotFoundException(100, "Source image not found");
    } else if (targetImage == null) {
        throw new NotFoundException(101, "Target image not found");
    } else if (threshold < 0 || threshold > 100) {
        throw new InvalidDataException(102, "Threshold should be between 0 and 100");
    } else {
        targets = new ArrayList<>();
        float targetArea = targetImage.getHeight() * targetImage.getWidth();
        for (int row = 0; row <= sourceImage.getHeight() - targetImage.getHeight(); row++) {
            for (int col = 0; col <= sourceImage.getWidth() - targetImage.getWidth(); col++) {
                int score = getScoreForSubSet(sourceImage, targetImage, row, col);
                float percentageMatch = ((score / targetArea) * 100);
                if (percentageMatch >= threshold) {
                    Target target = new Target();
                    target.setType(targetImage.getName());
                    target.setCoordinates(new Coordinates(row, col));
                    target.setPercentageMatch(percentageMatch);
                    targets.add(target);
                }
            }
        }
    }

    return targets;
}

2. getScoreForSubSet method for give sub matrix and position (row, col) of test data, it computes score, by comparing each pixel against targetImage.

private static int getScoreForSubSet(Image sourceImage, Image targetImage, int row, int col) {
    int score = 0;
    for (int targetRow = 0; targetRow < targetImage.getHeight(); targetRow++) {
        for (int targetCol = 0; targetCol < targetImage.getWidth(); targetCol++) {

            if (targetImage.getMatrix().get(targetRow).get(targetCol) == sourceImage.getMatrix().get(row + targetRow).get(col + targetCol)) {
                score++;
            }
        }
    }
    return score;
}