@SmallTest
public void testWordConfidences() {
final String inputText = "one two three";
final Bitmap bmp = getTextImage(inputText, 640, 480);
// Attempt to initialize the API.
final TessBaseAPI baseApi = new TessBaseAPI();
baseApi.init(TESSBASE_PATH, DEFAULT_LANGUAGE);
baseApi.setPageSegMode(TessBaseAPI.PageSegMode.PSM_SINGLE_BLOCK);
baseApi.setImage(bmp);
String text = baseApi.getUTF8Text();
assertNotNull("Recognized text is null.", text);
// Ensure that a mean confidence value is returned.
int conf = baseApi.meanConfidence();
boolean validConf = conf > 0 && conf <= 100;
assertTrue("Mean confidence value is incorrect.", validConf);
// Ensure that word confidence values are returned.
int numWords = text.split("\\s+").length;
int[] wordConf = baseApi.wordConfidences();
assertEquals("Found the wrong number of word confidence values.", numWords, wordConf.length);
for (int i = 0; i < wordConf.length; i++) {
boolean valid = 0 <= wordConf[i] && wordConf[i] <= 100;
assertTrue("Found an invalid word confidence value.", valid);
}
// Attempt to shut down the API.
baseApi.end();
bmp.recycle();
}
@SmallTest
public void testProgressValues() {
final String inputText = "hello";
final Bitmap bmp = getTextImage(inputText, 640, 480);
final Rect imageBounds = new Rect(0, 0, bmp.getWidth(), bmp.getHeight());
class Notifier implements ProgressNotifier {
public boolean receivedProgress = false;
@Override
public void onProgressValues(ProgressValues progressValues) {
receivedProgress = true;
testProgressValues(progressValues, imageBounds);
}
}
final Notifier notifier = new Notifier();
// Attempt to initialize the API.
final TessBaseAPI baseApi = new TessBaseAPI(notifier);
baseApi.init(TESSBASE_PATH, DEFAULT_LANGUAGE);
baseApi.setPageSegMode(TessBaseAPI.PageSegMode.PSM_SINGLE_LINE);
baseApi.setImage(bmp);
// Ensure that we receive a progress callback.
baseApi.getHOCRText(0);
assertTrue(notifier.receivedProgress);
// Attempt to shut down the API.
baseApi.end();
bmp.recycle();
}
@SmallTest
public void testEnd() {
final String inputText = "hello";
final Bitmap bmp = getTextImage(inputText, 640, 480);
// Attempt to initialize the API.
final TessBaseAPI baseApi = new TessBaseAPI();
baseApi.init(TESSBASE_PATH, DEFAULT_LANGUAGE);
baseApi.setPageSegMode(TessBaseAPI.PageSegMode.PSM_SINGLE_LINE);
baseApi.setImage(bmp);
// Ensure that getUTF8Text() fails after end() is called.
baseApi.end();
try {
baseApi.getUTF8Text();
fail("IllegalStateException not thrown");
} catch (IllegalStateException e) {
// Continue
} finally {
bmp.recycle();
}
// Ensure that reinitializing the API is successful.
boolean success = baseApi.init(TESSBASE_PATH, DEFAULT_LANGUAGE);
assertTrue("API failed to initialize after end()", success);
// Ensure setImage() does not throw an exception.
final Bitmap bmp2 = getTextImage(inputText, 640, 480);
baseApi.setImage(bmp2);
// Attempt to shut down the API.
baseApi.end();
bmp2.recycle();
}
@SmallTest
public void testSetRectangle() {
// Attempt to initialize the API.
final TessBaseAPI baseApi = new TessBaseAPI();
baseApi.init(TESSBASE_PATH, DEFAULT_LANGUAGE);
baseApi.setPageSegMode(TessBaseAPI.PageSegMode.PSM_SINGLE_CHAR);
final int width = 640;
final int height = 480;
final Bitmap bmp = Bitmap.createBitmap(width, height, Bitmap.Config.ARGB_8888);
final Paint paint = new Paint();
final Canvas canvas = new Canvas(bmp);
canvas.drawColor(Color.WHITE);
paint.setColor(Color.BLACK);
paint.setStyle(Style.FILL);
paint.setAntiAlias(true);
paint.setTextAlign(Align.CENTER);
paint.setTextSize(32.0f);
// Draw separate text on the left and right halves of the image.
final String leftInput = "A";
final String rightInput = "B";
canvas.drawText(leftInput, width / 4, height / 2, paint);
canvas.drawText(rightInput, width * 3 / 4, height / 2, paint);
baseApi.setVariable(TessBaseAPI.VAR_CHAR_WHITELIST, leftInput + rightInput);
baseApi.setImage(bmp);
// Ensure the result is correct for a rectangle on the left half of the image.
Rect left = new Rect(0, 0, width / 2, height);
baseApi.setRectangle(left);
String leftResult = baseApi.getUTF8Text();
assertEquals("Found incorrect text.", leftInput, leftResult);
// Ensure the result is correct for a rectangle on the right half of the image.
Rect right = new Rect(width / 2, 0, width, height);
baseApi.setRectangle(right);
String rightResult = baseApi.getUTF8Text();
assertEquals("Found incorrect text.", rightInput, rightResult);
// Attempt to shut down the API.
baseApi.end();
bmp.recycle();
}
@SmallTest
public void testSetVariable() {
final String inputText = "hello";
final Bitmap bmp = getTextImage(inputText, 640, 480);
// Attempt to initialize the API.
final TessBaseAPI baseApi = new TessBaseAPI();
baseApi.init(TESSBASE_PATH, DEFAULT_LANGUAGE);
baseApi.setPageSegMode(TessBaseAPI.PageSegMode.PSM_SINGLE_LINE);
// Ensure that setting the blacklist variable works.
final String blacklistedCharacter = inputText.substring(1, 2);
baseApi.setVariable(TessBaseAPI.VAR_CHAR_BLACKLIST, blacklistedCharacter);
baseApi.setImage(bmp);
final String outputText = baseApi.getUTF8Text();
assertFalse("Found a blacklisted character.", outputText.contains(blacklistedCharacter));
// Attempt to shut down the API.
baseApi.end();
bmp.recycle();
}
@SmallTest
public void testSetPageSegMode() {
// Attempt to initialize the API.
final TessBaseAPI baseApi = new TessBaseAPI();
baseApi.init(TESSBASE_PATH, DEFAULT_LANGUAGE);
// Check the default page segmentation mode.
assertEquals(
"Found unexpected default page segmentation mode.",
baseApi.getPageSegMode(),
DEFAULT_PAGE_SEG_MODE);
// Ensure that the page segmentation mode can be changed.
final int newPageSegMode = TessBaseAPI.PageSegMode.PSM_SINGLE_CHAR;
baseApi.setPageSegMode(newPageSegMode);
assertEquals(
"Found unexpected page segmentation mode.", baseApi.getPageSegMode(), newPageSegMode);
// Attempt to shut down the API.
baseApi.end();
}
@SmallTest
public void testClear() {
final String inputText = "hello";
final Bitmap bmp = getTextImage(inputText, 640, 480);
// Attempt to initialize the API.
final TessBaseAPI baseApi = new TessBaseAPI();
baseApi.init(TESSBASE_PATH, DEFAULT_LANGUAGE);
baseApi.setPageSegMode(TessBaseAPI.PageSegMode.PSM_SINGLE_LINE);
baseApi.setImage(bmp);
// Ensure that the getUTF8Text() operation fails after clear() is called.
baseApi.clear();
String text = baseApi.getUTF8Text();
assertNull("Received non-null result after clear().", text);
// Attempt to shut down the API.
baseApi.end();
bmp.recycle();
}
@SmallTest
public void testChoiceIterator() {
final String inputText = "hello";
final Bitmap bmp = TessBaseAPITest.getTextImage(inputText, 640, 480);
// Attempt to initialize the API.
final TessBaseAPI baseApi = new TessBaseAPI();
baseApi.init(TessBaseAPITest.TESSBASE_PATH, TessBaseAPITest.DEFAULT_LANGUAGE);
baseApi.setPageSegMode(TessBaseAPI.PageSegMode.PSM_SINGLE_LINE);
baseApi.setVariable(TessBaseAPI.VAR_SAVE_BLOB_CHOICES, TessBaseAPI.VAR_TRUE);
// Ensure that text is recognized.
baseApi.setImage(bmp);
String recognizedText = baseApi.getUTF8Text();
assertTrue("No recognized text found.", recognizedText != null && !recognizedText.equals(""));
// Iterate through the results.
ResultIterator iterator = baseApi.getResultIterator();
List<Pair<String, Double>> choicesAndConfidences;
iterator.begin();
do {
choicesAndConfidences = iterator.getChoicesAndConfidence(PageIteratorLevel.RIL_SYMBOL);
assertNotNull("Invalid result.", choicesAndConfidences);
for (Pair<String, Double> choiceAndConfidence : choicesAndConfidences) {
String choice = choiceAndConfidence.first;
Double conf = choiceAndConfidence.second;
assertTrue("No choice value found.", choice != null && !choice.equals(""));
assertTrue("Found an incorrect confidence value.", conf >= 0 && conf <= 100);
}
} while (iterator.next(PageIteratorLevel.RIL_SYMBOL));
iterator.delete();
assertNotNull("No ChoiceIterator values found.", choicesAndConfidences);
// Attempt to shut down the API.
baseApi.end();
bmp.recycle();
}
private void testGetHOCRText(String language, int ocrEngineMode) {
final String inputText = "hello";
final Bitmap bmp = getTextImage(inputText, 640, 480);
// Attempt to initialize the API.
final TessBaseAPI baseApi = new TessBaseAPI();
baseApi.init(TESSBASE_PATH, DEFAULT_LANGUAGE);
baseApi.setPageSegMode(TessBaseAPI.PageSegMode.PSM_SINGLE_LINE);
baseApi.setImage(bmp);
// Ensure that getHOCRText() produces a result.
final String hOcr = baseApi.getHOCRText(0);
assertNotNull("HOCR result not found.", hOcr);
assertTrue(hOcr.length() > 0);
final String outputText = Html.fromHtml(hOcr).toString().trim();
assertEquals(inputText, outputText);
// Attempt to shut down the API.
baseApi.end();
bmp.recycle();
}
@SmallTest
public void testProgressValues_setRectangle() {
class Notifier implements ProgressNotifier {
public boolean receivedProgress = false;
private Rect bounds;
public void reset(Rect bounds) {
this.bounds = bounds;
receivedProgress = false;
}
@Override
public void onProgressValues(ProgressValues progressValues) {
receivedProgress = true;
testProgressValues(progressValues, bounds);
}
}
final Notifier notifier = new Notifier();
final int width = 640;
final int height = 480;
final Bitmap bmp = Bitmap.createBitmap(width, height, Bitmap.Config.ARGB_8888);
final Paint paint = new Paint();
paint.setColor(Color.BLACK);
paint.setStyle(Style.FILL);
paint.setAntiAlias(true);
paint.setTextAlign(Align.CENTER);
paint.setTextSize(32.0f);
// Draw separate text on the left and right halves of the image.
final Canvas canvas = new Canvas(bmp);
canvas.drawColor(Color.WHITE);
final String leftInput = "A";
final String rightInput = "B";
canvas.drawText(leftInput, width / 4, height / 2, paint);
canvas.drawText(rightInput, width * 3 / 4, height / 2, paint);
// Attempt to initialize the API.
final TessBaseAPI baseApi = new TessBaseAPI(notifier);
baseApi.init(TESSBASE_PATH, DEFAULT_LANGUAGE);
baseApi.setPageSegMode(TessBaseAPI.PageSegMode.PSM_SINGLE_LINE);
baseApi.setVariable(TessBaseAPI.VAR_CHAR_WHITELIST, leftInput + rightInput);
baseApi.setImage(bmp);
// Attempt to restrict recognition to a sub-rectangle of the image.
final Rect left = new Rect(0, 0, width / 2, height);
baseApi.setRectangle(left);
notifier.reset(left);
// Ensure a progress callback is received.
baseApi.getHOCRText(0);
assertTrue(notifier.receivedProgress);
// Attempt to restrict recognition to a sub-rectangle of the image.
final Rect right = new Rect(width / 2 + 5, 7, width - 5, height - 7);
baseApi.setRectangle(right);
notifier.reset(right);
// Ensure a progress callback is received.
baseApi.getHOCRText(0);
assertTrue(notifier.receivedProgress);
// Attempt to shut down the API.
baseApi.end();
bmp.recycle();
}
private void testGetUTF8Text(String language, int ocrEngineMode) {
final String inputText = "hello";
final Bitmap bmp = getTextImage(inputText, 640, 480);
// Attempt to initialize the API.
final TessBaseAPI baseApi = new TessBaseAPI();
baseApi.init(TESSBASE_PATH, language, ocrEngineMode);
baseApi.setPageSegMode(TessBaseAPI.PageSegMode.PSM_SINGLE_LINE);
baseApi.setImage(bmp);
// Ensure that the result is correct.
final String outputText = baseApi.getUTF8Text();
assertEquals("\"" + outputText + "\" != \"" + inputText + "\"", inputText, outputText);
// Ensure getRegions() works.
final Pixa regions = baseApi.getRegions();
assertEquals("Found incorrect number of regions.", regions.size(), 1);
regions.recycle();
// Ensure getTextlines() works.
final Pixa textlines = baseApi.getTextlines();
assertEquals("Found incorrect number of textlines.", textlines.size(), 1);
textlines.recycle();
// Ensure getStrips() works.
final Pixa strips = baseApi.getStrips();
assertEquals("Found incorrect number of strips.", strips.size(), 1);
strips.recycle();
// Ensure getWords() works.
final Pixa words = baseApi.getWords();
assertEquals("Found incorrect number of words.", words.size(), 1);
words.recycle();
// Ensure getConnectedComponents() works.
final Pixa connectedComponents = baseApi.getConnectedComponents();
assertTrue("Connected components not found.", connectedComponents.size() > 0);
connectedComponents.recycle();
// Iterate through the results.
final ResultIterator iterator = baseApi.getResultIterator();
String lastUTF8Text;
float lastConfidence;
int[] lastBoundingBox;
Rect lastBoundingRect;
int count = 0;
iterator.begin();
do {
lastUTF8Text = iterator.getUTF8Text(PageIteratorLevel.RIL_WORD);
lastConfidence = iterator.confidence(PageIteratorLevel.RIL_WORD);
lastBoundingBox = iterator.getBoundingBox(PageIteratorLevel.RIL_WORD);
lastBoundingRect = iterator.getBoundingRect(PageIteratorLevel.RIL_WORD);
count++;
} while (iterator.next(PageIteratorLevel.RIL_WORD));
iterator.delete();
assertEquals("Found incorrect number of results.", count, 1);
assertEquals("Found an incorrect result.", lastUTF8Text, outputText);
assertTrue("Result was not high-confidence.", lastConfidence > 80);
assertTrue("Result bounding box not found.", lastBoundingBox[2] > 0 && lastBoundingBox[3] > 0);
boolean validBoundingRect =
lastBoundingRect.left < lastBoundingRect.right
&& lastBoundingRect.top < lastBoundingRect.bottom;
assertTrue("Result bounding box Rect is incorrect.", validBoundingRect);
// Attempt to shut down the API.
baseApi.end();
bmp.recycle();
}