Add flag for Page Segmentation Modes control (#1601)

* Add flag for Page Segmentation Modes control

I added an flag --psm for controlling PSM (Page Segmentation Modes) in Tesseract. The default option (3) gives me quite bad results. When I use 6, 11, or 12 for Bulgarian, it gives me much better OCR results. I haven't tested other languages yet, but I expect improvements as well if other mode is used.

* feat: add psm for rust parser

* fix: add psm to options

* fix: add default value of psm to 3

* fix: correct type of ocr oem

* fix(rust): use fatal! instead of exit

---------

Co-authored-by: Prateek Sunal <prtksunal@gmail.com>

docs/CHANGES.TXT

@@ -35,6 +35,7 @@
 - Fix: infinite loop in MP4 file type detector.
 - Improvement: Use Corrosion to build Rust code
 - Improvement: Ignore MXF Caption Essence Container version byte to enhance SRT subtitle extraction compatibility
+- New: Add tesseract page segmentation modes control with `--psm` flag
 
 0.94 (2021-12-14)
 -----------------

src/lib_ccx/ccx_common_option.c

@@ -72,6 +72,7 @@ void init_options(struct ccx_s_options *options)
  options->dvblang = NULL; // By default, autodetect DVB language
  options->ocrlang = NULL; // By default, autodetect .traineddata file
  options->ocr_oem = -1; // By default, OEM mode depends on the tesseract version
+ options->psm = 3; // Default PSM mode (3 is the default tesseract as well)
  options->ocr_quantmode = 1; // CCExtractor's internal
  options->mkvlang = NULL; // By default, all the languages are extracted
  options->ignore_pts_jumps = 1;

src/lib_ccx/ccx_common_option.h

@@ -147,6 +147,7 @@ struct ccx_s_options // Options from user parameters
  char *dvblang; // The name of the language stream for DVB
  const char *ocrlang; // The name of the .traineddata file to be loaded with tesseract
  int ocr_oem; // The Tesseract OEM mode, could be 0 (default), 1 or 2
+ int psm; // The Tesseract PSM mode, could be between 0 and 13. 3 is tesseract default
  int ocr_quantmode; // How to quantize the bitmap before passing to to tesseract (0=no quantization at all, 1=CCExtractor's internal)
  char *mkvlang; // The name of the language stream for MKV
  int analyze_video_stream; // If 1, the video stream will be processed even if we're using a different one for subtitles.

src/lib_ccx/ocr.c

@@ -177,6 +177,9 @@ void *init_ocr(int lang_index)
  &pars_values, 1, false);
  }
 
+ // set PSM mode
+ TessBaseAPISetPageSegMode(ctx->api, ccx_options.psm);
+
  free(pars_vec);
  free(pars_values);
 

src/lib_ccx/params.c

@@ -679,6 +679,23 @@ void print_usage(void)
  mprint(" Default value depends on the tesseract version linked :\n");
  mprint(" Tesseract v3 : default mode is 0,\n");
  mprint(" Tesseract v4 : default mode is 1.\n");
+ mprint(" --psm: Select the PSM mode for Tesseract.\n");
+ mprint(" Available Page segmentation modes:\n");
+ mprint(" 0 Orientation and script detection (OSD) only.\n");
+ mprint(" 1 Automatic page segmentation with OSD.\n");
+ mprint(" 2 Automatic page segmentation, but no OSD, or OCR.\n");
+ mprint(" 3 Fully automatic page segmentation, but no OSD. (Default)\n");
+ mprint(" 4 Assume a single column of text of variable sizes.\n");
+ mprint(" 5 Assume a single uniform block of vertically aligned text.\n");
+ mprint(" 6 Assume a single uniform block of text.\n");
+ mprint(" 7 Treat the image as a single text line.\n");
+ mprint(" 8 Treat the image as a single word.\n");
+ mprint(" 9 Treat the image as a single word in a circle.\n");
+ mprint(" 10 Treat the image as a single character.\n");
+ mprint(" 11 Sparse text. Find as much text as possible in no particular order.\n");
+ mprint(" 12 Sparse text with OSD.\n");
+ mprint(" 13 Raw line. Treat the image as a single text line,\n");
+ mprint(" bypassing hacks that are Tesseract-specific.\n");
  mprint(" --mkvlang: For MKV subtitles, select which language's caption\n");
  mprint(" stream will be processed. e.g. 'eng' for English.\n");
  mprint(" Language codes can be either the 3 letters bibliographic\n");
@@ -1696,6 +1713,27 @@ int parse_parameters(struct ccx_s_options *opt, int argc, char *argv[])
  fatal(EXIT_MALFORMED_PARAMETER, "--oem has no argument.\n");
  }
  }
+ if (strcmp(argv[i], "--psm") == 0)
+ {
+ if (i < argc - 1)
+ {
+ i++;
+
+ char *str = (char *)malloc(sizeof(argv[i]));
+ sprintf(str, "%s", argv[i]);
+ opt->psm = atoi(str);
+ if (opt->psm < 0 || opt->psm > 13)
+ {
+ fatal(EXIT_MALFORMED_PARAMETER, "--psm must be between 0 and 13\n");
+ }
+
+ continue;
+ }
+ else
+ {
+ fatal(EXIT_MALFORMED_PARAMETER, "--psm has no argument.\n");
+ }
+ }
  if (strcmp(argv[i], "--mkvlang") == 0)
  {
  if (i < argc - 1)

src/lib_ccx/params_dump.c

@@ -216,6 +216,8 @@ void params_dump(struct lib_ccx_ctx *ctx)
  mprint("Reduced color palette]\n");
  break;
  }
+
+ mprint("[Tesseract PSM: %d]\n", ccx_options.psm);
 }
 
 #define Y_N(cond) ((cond) ? "Yes" : "No")

src/rust/lib_ccxr/src/common/options.rs

@@ -455,6 +455,8 @@ pub struct Options {
  pub ocrlang: PathBuf,
  /// The Tesseract OEM mode, could be 0 (default), 1 or 2
  pub ocr_oem: i8,
+ /// The Tesseract PSM mode, could be between 0 and 13. 3 is tesseract default
+ pub psm: i32,
  /// How to quantize the bitmap before passing to to tesseract
  /// (0 = no quantization at all, 1 = CCExtractor's internal,
  /// 2 = reduce distinct color count in image for faster results.)
@@ -589,6 +591,7 @@ impl Default for Options {
  dvblang: Default::default(),
  ocrlang: Default::default(),
  ocr_oem: -1,
+ psm: 3,
  ocr_quantmode: 1,
  mkvlang: Default::default(),
  analyze_video_stream: Default::default(),

src/rust/src/args.rs

@@ -600,6 +600,25 @@ pub struct Args {
  /// Tesseract v4 : default mode is 1.
  #[arg(long, verbatim_doc_comment, value_name="mode", help_heading=OUTPUT_AFFECTING_OUTPUT_FILES)]
  pub oem: Option<u8>,
+ /// Select the PSM mode for Tesseract.
+ /// Available Page segmentation modes:
+ /// 0 Orientation and script detection (OSD) only.
+ /// 1 Automatic page segmentation with OSD.
+ /// 2 Automatic page segmentation, but no OSD, or OCR.
+ /// 3 Fully automatic page segmentation, but no OSD. (Default)
+ /// 4 Assume a single column of text of variable sizes.
+ /// 5 Assume a single uniform block of vertically aligned text.
+ /// 6 Assume a single uniform block of text.
+ /// 7 Treat the image as a single text line.
+ /// 8 Treat the image as a single word.
+ /// 9 Treat the image as a single word in a circle.
+ /// 10 Treat the image as a single character.
+ /// 11 Sparse text. Find as much text as possible in no particular order.
+ /// 12 Sparse text with OSD.
+ /// 13 Raw line. Treat the image as a single text line,
+ /// bypassing hacks that are Tesseract-specific.
+ #[arg(long, verbatim_doc_comment, value_name="mode", help_heading=OUTPUT_AFFECTING_OUTPUT_FILES)]
+ pub psm: Option<u8>,
  /// For MKV subtitles, select which language's caption
  /// stream will be processed. e.g. 'eng' for English.
  /// Language codes can be either the 3 letters bibliographic

src/rust/src/parser.rs

@@ -801,6 +801,16 @@ impl OptionsExt for Options {
  self.ocr_oem = *oem as _;
  }
 
+ if let Some(ref psm) = args.psm {
+ if !(0..=13).contains(psm) {
+ fatal!(
+ cause = ExitCause::MalformedParameter;
+ "--psm must be between 0 and 13"
+ );
+ }
+ self.psm = *psm as _;
+ }
+
  if let Some(ref lang) = args.mkvlang {
  self.mkvlang = Some(Language::from_str(lang.as_str()).unwrap());
  let str = lang.as_str();