Marmelator/ocr-service
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases 64 Wiki Activity

RED-7669: optimize OCR-module performance

Browse Source
This commit is contained in:
Kilian Schüttler 2023-12-12 15:27:00 +01:00 committed by Dominique Eifländer
parent 591c7d7fab
commit c06974ce69
4 changed files with 5 additions and 16 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

2
ocr-service-v1/ocr-service-processor/build.gradle.kts
View File

@ -20,7 +20,7 @@ dependencies {
api("org.apache.pdfbox:jbig2-imageio:3.0.4") api("org.apache.pdfbox:jbig2-imageio:3.0.4")
api("com.github.jai-imageio:jai-imageio-core:1.4.0") api("com.github.jai-imageio:jai-imageio-core:1.4.0")
api("com.github.jai-imageio:jai-imageio-jpeg2000:1.4.0") api("com.github.jai-imageio:jai-imageio-jpeg2000:1.4.0")
api("io.github.karols:hocr4j:0.1.2") api("io.github.karols:hocr4j:0.2.0")
api("com.amazonaws:aws-java-sdk-kms:1.12.440") api("com.amazonaws:aws-java-sdk-kms:1.12.440")
api("com.google.guava:guava:31.1-jre") api("com.google.guava:guava:31.1-jre")
api("com.iqser.red.commons:pdftron-logic-commons:2.20.0") api("com.iqser.red.commons:pdftron-logic-commons:2.20.0")

7
ocr-service-v1/ocr-service-processor/src/main/java/com/knecon/fforesight/service/ocr/processor/model/ExtractedOcrImage.java
View File

@ -58,11 +58,4 @@ public class ExtractedOcrImage implements OcrImage {
return affineTransform; return affineTransform;
} }
@Override
public int getOptimalPageSegmentationMode() {
return ITessAPI.TessPageSegMode.PSM_SINGLE_BLOCK;
}
} }

5
ocr-service-v1/ocr-service-processor/src/main/java/com/knecon/fforesight/service/ocr/processor/model/OcrImage.java
View File

@ -11,6 +11,7 @@ import lombok.SneakyThrows;
import net.sourceforge.lept4j.Leptonica1; import net.sourceforge.lept4j.Leptonica1;
import net.sourceforge.lept4j.Pix; import net.sourceforge.lept4j.Pix;
import net.sourceforge.lept4j.util.LeptUtils; import net.sourceforge.lept4j.util.LeptUtils;
import net.sourceforge.tess4j.ITessAPI;
public interface OcrImage { public interface OcrImage {
@ -84,7 +85,9 @@ public interface OcrImage {
* *
* @return The optimal page segmentation mode. * @return The optimal page segmentation mode.
*/ */
int getOptimalPageSegmentationMode(); // TODO: evaluate if PSM can be dynamically chosen to increase performance default int getOptimalPageSegmentationMode() {
return ITessAPI.TessPageSegMode.PSM_AUTO;
} // TODO: evaluate if PSM can be dynamically chosen to increase performance
/** /**

7
ocr-service-v1/ocr-service-processor/src/main/java/com/knecon/fforesight/service/ocr/processor/model/RenderedPageOcrImage.java
View File

@ -28,13 +28,6 @@ public class RenderedPageOcrImage implements OcrImage {
int rotationDegrees; int rotationDegrees;
@Override
public int getOptimalPageSegmentationMode() {
return ITessAPI.TessPageSegMode.PSM_SINGLE_BLOCK;
}
@Override @Override
public AffineTransform getImageCTM() { public AffineTransform getImageCTM() {