Word and character segmentation directly in run-length compressed handwritten document images

Amarnath R; Mohammed Javed; P. Nagabhushan

REVIEW 4 major objections 5 minor 5 references

Word and character segmentation directly in run-length compressed handwritten document images

T0 review · 4 major / 5 minor · reviewed 2026-08-14 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read Handwritten word and character segmentation can be done directly on run-length compressed document images without decompressing to pixels.

desk verdict A genuinely new compressed-domain word/character segmentation idea, undermined by a mislabeled 'min-cut' heuristic and an evaluation too weak to support the accuracy claims. read the letter →

arxiv 1909.05146 v1 pith:RPKBNUDS submitted 2019-08-18 cs.CV

Amarnath R , P. Nagabhushan , Mohammed Javed This is my paper · ORCID

classification cs.CV

keywords wordsegmentationcharacterrun-lengthencodingcompresseddomaindocumentimageanalysishandwrittenimagesconnectedcomponentsmin-cutgraphICDAR13

verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

This paper sets out to prove that word and character segmentation, a step normally run on decompressed pixel images, can be performed directly on the run-length encoded representation of a handwritten document image. The motivation is practical: if segmentation can run on the compressed form, document analysis pipelines avoid repeated decompression and recompression, saving both computation time and memory. The authors trace the horizontal spread of foreground runs to build connected components, split words by the large gaps between those components, and separate touching characters with a min-cut procedure on band-wise run patterns. They report 90.29% word segmentation accuracy and 86.23% cascaded character segmentation accuracy on the ICDAR13 English handwritten dataset, with lower accuracy when the same method is applied to non-English scripts.

What carries the argument

The machinery is the run-length representation of the image, together with a cumulative-run computation that turns each foreground run into a component spread [xmin, xmax]. Word segmentation uses an empirically chosen threshold on the gaps between these spreads to separate inter-word gaps from intra-word gaps. Character segmentation restricts attention to a region of interest, splits it into three horizontal bands, and performs a logical OR of the top and bottom bands; the gaps in the resulting pattern are treated as minimum cut edges where adjacent characters separate. Over- and under-segmentation are then repaired by inserting or deleting separator points according to component lengths relative to the average.

What would settle it

Run the OR-of-bands character segmentation on a set of English handwritten words whose ground truth includes many full-height touching letters, such as tightly cursive 'th' or 'll' joins; if the OR pattern shows no gap at a large share of those true boundaries, the central assumption fails. A simpler check is to measure the cascaded character accuracy on a subset of ICDAR13 marked for heavy cursive versus a subset with separated letters; the method should perform systematically worse on the cursive subset.

Watch

Extended reading notes

Core claim

The core claim is that the geometry of handwriting segmentation is visible directly in run-length encoded data: connected components can be established purely from the horizontal spreads of foreground runs, inter-word boundaries are the large gaps between those components, and within a word the touching of adjacent characters can be resolved by a logical OR of the top and bottom bands of the word's region of interest. The authors show that this is enough to reach 90.29% word accuracy and 86.23% cascaded character accuracy on ICDAR13, and they argue that working in the compressed domain costs O(m×n) time in the worst case, the same order as reading the compressed data itself, while pixel-domain processing would cost O(h×w). The character-segmentation step is explicitly built for English handwriting, where connections between neighboring letters occur mostly in the mid-bottom region.

Load-bearing premise

The character segmentation step assumes that adjacent letters in English handwriting overlap only in the mid-bottom region, so that overlaying the top and bottom bands reveals the cuts; if letters touch across their full height or the writing is heavily cursive, the OR pattern will not expose a separator and characters will be under-segmented.

Editorial extensions

If this is right

OCR-style pipelines that already receive compressed scans can segment words and characters without decompressing to a pixel grid, removing the buffer and time cost of repeated decode-encode cycles.
Because the word boundary test is just a gap-length threshold, the method transfers to any script whose word spacing is consistently larger than intra-word spacing; the accuracy figures on Kannada, Bangla, Oriya, and Persian show the transfer is partial.
Character accuracy is cascaded: errors in word segmentation propagate directly into character segmentation, so a 90.29% word accuracy caps the ceiling for the 86.23% character accuracy.
The method requires no training and no pixel classification, only run-length data, which makes it a candidate for memory-constrained embedded document scanners.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

A natural next step is to replace the empirically fixed gap threshold with a data-driven bimodality test on the component-gap histogram; because the paper already computes average gaps, a statistical split point would make the method parameter-free and easier to port across scripts.
The OR-of-bands procedure encodes a general principle: in any script where character joining is confined to a narrow vertical zone, logical combinations of band-wise run patterns will expose separators; this suggests a testable extension for Arabic, Devanagari, or Chinese by first locating the joining zone automatically.
Since the spread-based features are purely horizontal, the method's sensitivity to skew could be quantified directly: rotating a test line by small angles and measuring when word accuracy drops would give an explicit skew tolerance, which the paper leaves outside its scope.

Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Desk Editor's Note

read the letter

Mohammed,

This is an incremental but genuinely new extension: the authors take their compressed-domain line segmentation and push it to word and character level for handwritten documents. Nobody else seems to have done that directly on run-length encoded data, so the niche is real. They also deserve credit for reporting performance on multiple non-English scripts and admitting the method degrades there.

The central approach is plausible: find connected components from foreground runs, use larger gaps to split words, and for character cuts, OR the top and bottom horizontal bands of the ROI and use the first x-gap in that OR. The complexity argument—O(mn) on the compressed representation versus O(hw) on the pixel grid—is fine and could matter for large archives.

The soft spots are substantial. The 'min-cut graph' is not a graph cut; it is a band-OR heuristic. It only works when adjacent characters touch in the mid-bottom region and leave a gap in the top and bottom bands. The paper asserts this is typical for English handwriting but supplies no measurement of how often ICDAR13 words satisfy it. In slanted or cursive hands, characters touch across the full height and the OR band is continuous, so the method silently fails. The reported 86.23% cascaded accuracy is therefore not anchored to a measured property; it is a fitted value on a specific dataset. Thresholds for the inter-word gap and ROI are hand-set, there are no error bars, no comparison against any decompress-then-segment baseline, and no code or data. I also could not follow equations (2)–(6): CR(i) is defined as a row sum, then used as if it were per-column coordinates, and the coordinate mapping is not reproducible.

Who benefits: researchers working on compressed document image analysis, especially those building on Javed-style RLE processing. It is a useful existence proof, not a reliable method yet. I would send it to peer review if the venue tolerates heavy revision; the idea is worth one round of serious refereeing. A desk reject would also be defensible given the evaluation, but I think the uniqueness of the compressed-domain claim tips it toward review.

Referee Report

4 major / 5 minor

Summary. The paper proposes methods for word and character segmentation performed directly on run-length encoded (RLE) handwritten document images. Word segmentation identifies connected components of foreground runs and thresholds inter-component gaps to group components into words. Character segmentation restricts analysis to a region of interest (ROI), divides it into top, middle, and bottom horizontal bands, performs a logical OR of the top and bottom bands to expose cut points between touching characters (described as a min-cut graph concept), and then applies insertion/deletion rules based on component lengths to address over- and under-segmentation. Experiments on ICDAR13 and several non-English datasets report 90.29% word segmentation accuracy and 86.23% cascaded character segmentation accuracy for English, with lower accuracies for Kannada, Oriya, Bangla, and Persian. The paper also claims lower computational complexity relative to pixel-domain processing and presents qualitative results on machine-printed images.

Significance. If the central claim holds, namely that word and character segmentation can be performed accurately and efficiently in the compressed domain, the work would be a useful contribution to document image analysis, especially for applications that need to avoid repeated decompression and recompression. The idea of exploiting run-length structure for segmentation is attractive, and the reported accuracies on ICDAR13 are non-trivial. The paper's strongest conceptual asset is the complexity argument for compressed-domain processing, which could motivate further work in this direction. However, the paper does not provide a direct comparison with existing segmentation methods, does not report sample sizes or error bars, does not make code or data available, and does not validate the key assumption about how adjacent characters touch. The thresholds are empirically selected from the test data, so the reported numbers are in-sample fits rather than predictions. Consequently, the significance of the claimed results cannot be fully assessed from the manuscript as it stands.

major comments (4)

[Section 3, Algorithm Step 4 and Figure 14] The algorithm described as a 'min-cut graph concept' is actually a band-OR heuristic: no graph, cut edges, or edge weights are defined, and the separation point is simply the first x-interval in which the logical OR of the top and bottom bands has no foreground. This mischaracterization affects the claimed novelty of using a graph-based min-cut. More importantly, the method's key assumption that adjacent characters connect only in the mid-bottom region is unvalidated: the paper does not report how often ICDAR13 character pairs satisfy this condition, so the 86.23% cascaded accuracy is not anchored to the assumption actually holding. Cursive or full-height/slanted joins would make the OR band continuous across the boundary and the method would fail. Provide statistics on touching-character geometry and compare against a pixel-domain segmentation baseline.
[Section 2.2, Tables 2 and 4; Section 3.2, Equation (8)] No sample sizes, number of test images, words, characters, error bars, or statistical tests are reported, and the matching criterion behind equations (7) and (8) is unspecified. The 90.29% word and 86.23% character accuracies are raw in-sample match rates on an unspecified subset of ICDAR13. Without a comparison to ICDAR13 contest participants or other published segmentation methods, the central claim that compressed-domain processing is competitive cannot be assessed. Additionally, the character accuracy is cascaded after word segmentation, so errors propagate; a standalone evaluation on ground-truth word segments is needed to isolate character-segmentation performance.
[Section 2, Equations (2)-(6) and Section 2.1 complexity analysis] Equations (2)-(6) are internally inconsistent and appear to treat the run-length representation as a binary image. Equation (2) defines CR(i) as a row sum over RLE(i,j) but does not define what RLE(i,j) represents for a run-length encoding, while equation (6) simultaneously uses RLE(i,j) as a binary pixel value and as a run-length term. As a result, the O(m x n) complexity claim is unsupported: if m and n are uncompressed image dimensions, the algorithm is not actually operating in the compressed domain; if they are compressed dimensions, the equations and complexity bound need a proper RLE formulation. This directly affects the paper's central premise that processing directly in compressed data saves time and memory.
[Section 2.1 Algorithm Step 3 vs. Sections 2 and 3.1] Threshold selection is described inconsistently: Section 2 states a threshold is 'empirically chosen,' Algorithm Step 3 says 'Compute average gaps to automatically select a threshold value,' and Section 3.1 uses an empirically set ROI threshold t. If thresholds are computed from the test data themselves, the reported accuracies are in-sample fits rather than predictions. The paper must specify a training/test protocol or provide a sensitivity analysis over threshold values to show that the results are not due to dataset-specific tuning.

minor comments (5)

[Section 4] Section 4 presents only two figures for machine-printed images with no quantitative evaluation; either provide numerical results or remove the claim that the model has been tested on machine-printed document images.
[Abstract and Conclusion] The abstract and conclusion state that a 'min-cut graph concept' is used, but the body of the paper does not define a graph; please align the terminology with the actual OR-band procedure described in Section 3.
[Section 2] The wording 'spread of every character' is misleading because the algorithm computes connected components of foreground runs, which may merge multiple characters; consider rewording to 'spread of connected components.'
[Equations (7) and (8)] The 'one to one matching' used in equations (7) and (8) is not defined; specify the matching rule, including whether partial overlaps count and how ground-truth entities are matched to detected segments.
[References and Datasets] Reference [Javed et al, 2016] and the later citation to 'Javed et al (Ph.D. Thesis), 2016' appear to refer to the same work; also, the datasets 'PBOK (Section 1.6.1)' and 'MARG dataset' are not described in the paper, so the reader cannot reproduce the experiments.

Circularity Check

0 steps flagged · score 2.0 of 10

No circular derivation is present: reported accuracies rest on empirical thresholds and a heuristic OR-gap procedure, not on equations that reduce to the paper's own outputs.

full rationale

The paper does not contain a load-bearing circular step that the instructions require before flagging circularity. Word segmentation (Section 2) builds connected components from RLE runs using Eqs. (2)-(4) and separates words by an inter-word gap threshold. The paper says 'a threshold value is fixed empirically to distinguish words from intra-word gaps' and its Algorithm Step 3 says 'Compute average gaps to automatically select a threshold value.' This means the threshold is fitted to the data, so the ICDAR13 accuracy in Table 2 is an empirical evaluation rather than a parameter-free prediction. However, the accuracy is not derived from the threshold by construction, and the threshold is not a renamed form of the ground-truth output. Character segmentation (Section 3) divides the ROI into three horizontal blocks, performs a logical OR of the top and bottom portions, and treats gaps in the OR pattern as cut edges. The stated assumption that 'the connection between the adjacent characters mostly occurs in the mid-bottom regions' is a modeling assumption that limits robustness, but the algorithm output is not fed back into that assumption, and no equation reduces the reported 86.23% accuracy to the assumption itself. The use of the term 'min-cut graph concept' is an overstatement because no actual graph-cut objective is defined; the procedure is an OR-gap heuristic. That is a correctness or naming concern, not circularity. The self-citations to the authors' prior text-line segmentation work are used as motivation and as pipeline input, not as a uniqueness theorem or as the source of the word and character segmentation results. Thus no specific reduction (Eq. X = Eq. Y by construction, or fitted parameter renamed as prediction) can be exhibited, and the honest finding is no significant circularity. The score of 2 reflects only the minor non-circular caveats of empirically chosen thresholds and non-load-bearing self-citations.

Assumptions & free parameters 3 free parameters · 3 assumptions · 0 invented entities

The central methods rest on two empirically chosen thresholds and two domain assumptions about handwriting geometry. No new physical or mathematical entities are introduced.

free parameters (3)

Inter-word gap threshold = not reported
Empirically chosen threshold used to distinguish inter-word gaps from intra-word gaps; Section 2 states 'a threshold value is fixed empirically'.
ROI threshold t = not reported
Threshold parameter for selecting the region of interest in character segmentation; Section 3.1 lists t as input but no value or selection rule is given.
Over/under-segmentation length thresholds = not reported
Component length thresholds for insertion and deletion operations; Section 3 says decisions are 'entirely based on the size of every connected component', but the exact rule is not specified.

assumptions (3)

domain assumption Inter-word gaps are larger than intra-word gaps in handwritten text-lines.
The word segmentation method relies on this spacing property; Section 2: 'The inter-word gaps are perceivably large when compared to the intra-word gaps.'
domain assumption In English handwriting, adjacent characters connect mostly in the mid-bottom region.
Character segmentation uses OR of top and bottom bands to find cuts; Section 3 states this explicitly.
domain assumption Run-length encoded foreground runs can be aggregated into connected components that correspond to characters or words.
The method builds connected components from the spans of foreground runs; no proof that this mapping is reliable under skew or noise is given.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Word and character segmentation directly in run-length compressed handwritten document images." pith.science (2026). https://pith.science/paper/RPKBNUDS

@misc{pith2026190905146,
  author       = {Pith},
  title        = {Pith review of: Word and character segmentation directly in run-length compressed handwritten document images},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/RPKBNUDS}},
  note         = {Machine review of arXiv:1909.05146}
}

read the original abstract

From the literature, it is demonstrated that performing text-line segmentation directly in the run-length compressed handwritten document images significantly reduces the computational time and memory space. In this paper, we investigate the issues of word and character segmentation directly on the run-length compressed document images. Primarily, the spreads of the characters are intelligently extracted from the foreground runs of the compressed data and subsequently connected components are established. The spacing between the connected components would be larger between the adjacent words when compared to that of intra-words. With this knowledge, a threshold is empirically chosen for inter-word separation. Every connected component within a word is further analysed for character segmentation. Here, min-cut graph concept is used for separating the touching characters. Over-segmentation and under-segmentation issues are addressed by insertion and deletion operations respectively. The approach has been developed particularly for compressed handwritten English document images. However, the model has been tested on non-English document images.

Figures

Figures reproduced from arXiv: 1909.05146 by the authors.

**Figure 2.** Connected components based on the spread of characters [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗

**Figure 3.** illustrates the process with an example. Here we have two connected components, namely c1 and c2. A small gap can be spotted in column number 18. Thus, we could trace the spread of every character by using equation 1. Further, [PITH_FULL_IMAGE:figures/full_fig_p004_3.png] view at source ↗

Figures from the paper (16 more)

**Figure 4.** Figure 4: Spread of foreground runs for all components [PITH_FULL_IMAGE:figures/full_fig_p004_4.png]

**Figure 5.** Figure 5: Sample input and the spread of characters along x [PITH_FULL_IMAGE:figures/full_fig_p005_5.png]

**Figure 6.** Figure 6: Lengths of the word and inter-word spacing The final step is to find the coordinates in RLE with respect to traced spatial locations. This can be formulated as [PITH_FULL_IMAGE:figures/full_fig_p005_6.png]

**Figure 7.** Figure 7: Results of word segmentation in the uncompressed version [PITH_FULL_IMAGE:figures/full_fig_p006_7.png]

**Figure 8.** Figure 8: Comparative computational time analysis of CDP and PDP for ICDAR13 dataset [PITH_FULL_IMAGE:figures/full_fig_p007_8.png]

**Figure 9.** Figure 9: A sample word image [PITH_FULL_IMAGE:figures/full_fig_p009_9.png]

**Figure 10.** Figure 10: Segmented components The next step is to find the region of interest (ROI) for analysis. Here, a threshold value is set empirically for the selection of ROI [PITH_FULL_IMAGE:figures/full_fig_p009_10.png]

**Figure 11.** Figure 11: Region of interest for character separation [PITH_FULL_IMAGE:figures/full_fig_p010_11.png]

**Figure 12.** Figure 12: Divided portions for analysis (a) Top portion (b) Middle portion (c) Bottom portion [PITH_FULL_IMAGE:figures/full_fig_p010_12.png]

**Figure 13.** Figure 13: Frequency of bands in every divided portion [PITH_FULL_IMAGE:figures/full_fig_p010_13.png]

**Figure 14.** Figure 14: Character segmentation results. The final step is to identify over-segmentation and under-segmentation regions. This is entirely based on the size of every connected component [PITH_FULL_IMAGE:figures/full_fig_p011_14.png]

**Figure 15.** Figure 15: Result of character segmentation [PITH_FULL_IMAGE:figures/full_fig_p012_15.png]

**Figure 16.** Figure 16: Results of character segmentation in the uncompressed version [PITH_FULL_IMAGE:figures/full_fig_p012_16.png]

**Figure 17.** Figure 17: Comparative computational time analysis of CDP and PDP for ICDAR13 dataset [PITH_FULL_IMAGE:figures/full_fig_p013_17.png]

**Figure 18.** Figure 18: Word Segmentation (Section 2): MARG dataset [PITH_FULL_IMAGE:figures/full_fig_p015_18.png]

**Figure 19.** Figure 19: Character Segmentation (Section 3): MARG dataset [PITH_FULL_IMAGE:figures/full_fig_p016_19.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

5 extracted references · 5 canonical work pages

[1]

With the advent of imaging devices, image based documents were born which were easy to store and transmission

Introduction Paper documents are being used as the primary source of communication and preserving information. With the advent of imaging devices, image based documents were born which were easy to store and transmission. Generally, a digital document image occupies a large storage space and huge bandwidth for transmission. Therefore, image compression al...

work page 2006
[2]

The overall idea is to find the inter-word gaps in a text -line directly in the run -length represented handwritten document image

Word segmentation In this section, word segmentation method is explained in detail. The overall idea is to find the inter-word gaps in a text -line directly in the run -length represented handwritten document image. The outline of the procedure is as follows. The f irst step is to identify the spread of every character in a text -line. This is achieved by...

work page 2016
[3]

Character segmentation In this section, we extend the previously discussed word segmentation technique to accomplish character segmentation specifically focusing on English handwritten document images in the compressed format . Here, the notion is to further analyse and refine ev ery connected component in a text -line for locating the character spacing, ...

work page
[4]

Two models are presented in Figure 18, and 19 respectively

Experimental results on Compressed Machine Printed Document Images This section presents some of the experimental results of word segmentation and character segmentation that have been carried out on compressed machine printed document images. Two models are presented in Figure 18, and 19 respectively. Figure 18: Word Segmentation (Section 2): MARG datase...

work page 2016
[5]

The word segmentation approach is based on the inter-word gaps whereas the character segmentation is based on the intra -word spacing

Conclusion In this article, we have proposed techniques for segmenting the compressed handwritten document images into words and characters. The word segmentation approach is based on the inter-word gaps whereas the character segmentation is based on the intra -word spacing. We have employed graph-based min-cut concept for segmenting the characters. Exper...

work page 2017

Pith tools

Reviewed August 14, 2026 · model on record in the stance chip above.

[1] [1]

With the advent of imaging devices, image based documents were born which were easy to store and transmission

Introduction Paper documents are being used as the primary source of communication and preserving information. With the advent of imaging devices, image based documents were born which were easy to store and transmission. Generally, a digital document image occupies a large storage space and huge bandwidth for transmission. Therefore, image compression al...

work page 2006

[2] [2]

The overall idea is to find the inter-word gaps in a text -line directly in the run -length represented handwritten document image

Word segmentation In this section, word segmentation method is explained in detail. The overall idea is to find the inter-word gaps in a text -line directly in the run -length represented handwritten document image. The outline of the procedure is as follows. The f irst step is to identify the spread of every character in a text -line. This is achieved by...

work page 2016

[3] [3]

Character segmentation In this section, we extend the previously discussed word segmentation technique to accomplish character segmentation specifically focusing on English handwritten document images in the compressed format . Here, the notion is to further analyse and refine ev ery connected component in a text -line for locating the character spacing, ...

work page

[4] [4]

Two models are presented in Figure 18, and 19 respectively

Experimental results on Compressed Machine Printed Document Images This section presents some of the experimental results of word segmentation and character segmentation that have been carried out on compressed machine printed document images. Two models are presented in Figure 18, and 19 respectively. Figure 18: Word Segmentation (Section 2): MARG datase...

work page 2016

[5] [5]

The word segmentation approach is based on the inter-word gaps whereas the character segmentation is based on the intra -word spacing

Conclusion In this article, we have proposed techniques for segmenting the compressed handwritten document images into words and characters. The word segmentation approach is based on the inter-word gaps whereas the character segmentation is based on the intra -word spacing. We have employed graph-based min-cut concept for segmenting the characters. Exper...

work page 2017