pith. sign in

arxiv: 2604.25464 · v2 · submitted 2026-04-28 · 💻 cs.CV

Image Compression with Bubble-Aware Frame Rate Adaptation for Energy-Efficient Video Capsule Endoscopy

Oliver Bause , J\"org Gamerdinger , Julia Werner , Oliver Bringmann This is my paper

Pith reviewed 2026-05-07 16:39 UTC · model grok-4.3

classification 💻 cs.CV
keywords video capsule endoscopyimage compressionenergy efficiencyframe rate adaptationbubble detectiongastrointestinal imagingmedical devices
0
0 comments X

The pith

A compression pipeline for video capsule endoscopy uses its own artifacts to flag bubble-obscured frames and drops the frame rate during those intervals, cutting energy use by up to 40 percent while preserving diagnostic quality.

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

The paper sets out to show that a single compression step can both shrink the data volume transmitted from a battery-limited capsule and reveal which frames carry little diagnostic value because bubbles block the view. By lowering the acquisition rate only in those low-visibility stretches, the method saves additional power without requiring separate bubble detectors or risking missed anomalies. A sympathetic reader would care because video capsule endoscopy is limited today by short battery life that often prevents full examination of the small intestine; extending runtime could make the procedure more complete and reliable.

Core claim

The authors demonstrate a compression scheme that delivers a ratio of 5.748 at 40.3 dB PSNR on Kvasir-Capsule and Galar data, yielding a mean 20.58 percent reduction in whole-system energy on a RISC-V platform. The same compression process supplies an indicator of bubble presence that drives a dynamic frame-rate adaptation, producing up to 40 percent further energy savings while the quality metric indicates negligible loss of visual information.

What carries the argument

Bubble-aware frame rate adaptation that exploits compression artifacts to identify low-diagnostic-value frames without any additional image analysis step.

If this is right

  • Compression alone reduces mean system energy by 20.58 percent.
  • Bubble-aware adaptation adds up to 40 percent energy reduction.
  • Image quality remains high at 40.3 dB PSNR, indicating negligible diagnostic loss.
  • The approach runs on a RISC-V platform and was tested on standard VCE datasets.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • Longer battery life could allow capsules to image the full length of the small intestine more consistently.
  • The same artifact-based detection idea might apply to other battery-constrained wireless cameras where obstructions create low-value frames.
  • A practical next test would be to quantify whether the adaptive mode changes the time to first detection of bleeding or ulcers in real procedures.

Load-bearing premise

Frames that the compression process marks as low-visibility because of bubbles truly contain negligible diagnostic information, so lowering the frame rate in those intervals will not cause clinically relevant anomalies to be missed.

What would settle it

A side-by-side clinical comparison on the same patients that measures the anomaly detection rate when the adaptive low-frame-rate periods are used versus when every frame is captured at the full rate.

Figures

Figures reproduced from arXiv: 2604.25464 by J\"org Gamerdinger, Julia Werner, Oliver Bause, Oliver Bringmann.

Figure 1
Figure 1. Figure 1: Proposed method of the RAW image compressing view at source ↗
Figure 2
Figure 2. Figure 2: Example images from Kvasir-Caspule and Galar dataset view at source ↗
Figure 3
Figure 3. Figure 3: Relation between the CR of the Bayer images and the view at source ↗
Figure 5
Figure 5. Figure 5: Inconsistent frame classification of Kvasir-Capsule with view at source ↗
Figure 6
Figure 6. Figure 6: Distribution of the images with a pathology label of view at source ↗
Figure 8
Figure 8. Figure 8: Simulation of energy consumption of the 80 Galar studies with different CR Threshold and fps reduction combinations. view at source ↗
read the original abstract

Video Capsule Endoscopy (VCE) is a promising method for improving the medical examination of the small intestine in the gastrointestinal tract. A key challenge is their limited size, resulting in a short battery lifetime which conflicts with high energy consumption for image capturing and transmission to an on-body device. Thus, we propose an image compression pipeline that substantially reduces the transmitted data while preserving diagnostic image quality. Furthermore, we exploit characteristics of the compression process to identify frames with low diagnostic value mainly caused by bubbles, without requiring additional image analysis. For low-visibility frames, a dynamic bubble-aware frame rate adaptation strategy reduces image acquisition and transmission during these phases while preserving sensitivity to potential anomalies. The proposed compression and frame rate adaptation are evaluated on a RISC-V platform using the Kvasir-Capsule and Galar datasets. The compression method achieves a compression ratio of 5.748 (82.6%) at a peak signal-to-noise ratio of 40.3 dB, indicating negligible loss of visual quality. The compression accomplished a mean energy reduction of the whole system by 20.58%. Additionally, the proposed bubble-aware frame rate adaptation reduced the energy consumption by up to 40%. These results demonstrate the potential of our method to increase the applicability of VCE.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

1 major / 1 minor

Summary. The paper proposes an image compression pipeline for Video Capsule Endoscopy (VCE) that reduces transmitted data while preserving diagnostic quality, and exploits compression artifacts to identify low-visibility frames (primarily due to bubbles) without additional analysis. A dynamic bubble-aware frame rate adaptation then lowers acquisition and transmission rates during these intervals. Evaluations on the Kvasir-Capsule and Galar datasets using a RISC-V platform report a compression ratio of 5.748 (82.6%) at 40.3 dB PSNR, 20.58% mean system energy reduction from compression, and up to 40% further reduction from frame-rate adaptation.

Significance. If the safety assumption holds, the work could extend VCE battery life by 20-40% while maintaining diagnostic utility, addressing a core hardware constraint. Strengths include concrete, reproducible empirical measurements (compression ratio, PSNR, energy figures) on named public datasets and a real RISC-V platform rather than simulation alone; the parameter-free reuse of compression artifacts for bubble detection is an efficient design choice.

major comments (1)
  1. [Evaluation] Evaluation section (Kvasir-Capsule and Galar results): The claim that low-visibility frames identified via compression artifacts contain negligible diagnostic information and can safely have frame rate reduced without missing anomalies is load-bearing for the 40% energy-reduction result, yet no quantitative validation is provided. No sensitivity/specificity analysis, expert-labeled anomaly masks, or comparison of flagged vs. unflagged frames for clinically relevant events is reported, leaving open the possibility that the energy savings trade off diagnostic safety.
minor comments (1)
  1. [Method] The description of how compression artifacts are thresholded to flag bubbles could be expanded with pseudocode or a precise decision rule to improve reproducibility.

Simulated Author's Rebuttal

1 responses · 0 unresolved

We thank the referee for their insightful comments on our manuscript. We provide a point-by-point response to the major comment below and outline the revisions we will make.

read point-by-point responses

  1. Referee: [Evaluation] Evaluation section (Kvasir-Capsule and Galar results): The claim that low-visibility frames identified via compression artifacts contain negligible diagnostic information and can safely have frame rate reduced without missing anomalies is load-bearing for the 40% energy-reduction result, yet no quantitative validation is provided. No sensitivity/specificity analysis, expert-labeled anomaly masks, or comparison of flagged vs. unflagged frames for clinically relevant events is reported, leaving open the possibility that the energy savings trade off diagnostic safety.

    Authors: The referee correctly identifies that our evaluation does not include a sensitivity or specificity analysis using expert-labeled anomaly masks for the frame-rate adaptation component. This omission stems from the fact that the public datasets employed (Kvasir-Capsule and Galar) lack the detailed per-frame clinical annotations required for such a study. Instead, the bubble-aware adaptation is motivated by prior medical literature indicating that bubbles in VCE images substantially reduce visibility and diagnostic content. By detecting these frames via compression artifacts, we reduce acquisition and transmission rates only during periods of inherently low diagnostic value. In the revised manuscript, we will add a dedicated paragraph in the discussion section to elaborate on these assumptions, reference relevant VCE studies on bubble effects, and acknowledge that comprehensive clinical validation would be necessary to fully confirm the safety of the energy savings. This revision will better contextualize the reported up to 40% additional energy reduction. revision: yes

Circularity Check

0 steps flagged

No circularity: empirical measurements on external datasets with no self-referential derivations

full rationale

The paper reports measured compression ratios, PSNR values, and energy savings from running the proposed pipeline on the Kvasir-Capsule and Galar datasets. The bubble-aware frame-rate logic is described as exploiting existing compression artifacts rather than introducing a fitted model or self-defined prediction. No equations, uniqueness theorems, or ansatzes are presented that reduce by construction to the paper's own inputs or prior self-citations. The central claims remain falsifiable external measurements; the noted limitation (lack of expert-labeled anomaly checks) is a correctness concern, not a circularity in the derivation chain.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 0 invented entities

No free parameters, axioms, or invented entities are introduced; the work is an empirical engineering pipeline evaluated on existing datasets.

pith-pipeline@v0.9.0 · 5534 in / 1106 out tokens · 34100 ms · 2026-05-07T16:39:52.248875+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

23 extracted references · 23 canonical work pages

  1. [1]

    Wireless capsule endoscopy,

    G. Iddan, G. Meron, A. Glukhovsky, and P. Swain, “Wireless capsule endoscopy,” Nature, vol. 405, no. 6785, pp. 417–417, 2000

    work page 2000

  2. [2]

    Wireless capsule endoscopy of the small bowel: development, testing, and first human trials,

    P. Swain, G. J. Iddan, G. Meron, and A. Glukhovsky, “Wireless capsule endoscopy of the small bowel: development, testing, and first human trials,” in Biomonitoring and Endoscopy Technologies, vol. 4158. SPIE, 2001, pp. 19–23

    work page 2001

  3. [3]

    A prospective trial comparing small bowel radiographs and video capsule endoscopy for suspected small bowel disease,

    G. Costamagna, S. K. Shah, M. E. Riccioni, F. Foschia, M. Mutignani, V . Perri, A. Vecchioli, M. G. Brizi, A. Picciocchi, and P. Marano, “A prospective trial comparing small bowel radiographs and video capsule endoscopy for suspected small bowel disease,” Gastroenterology, vol. 123, no. 4, pp. 999–1005, 2002

    work page 2002

  4. [4]

    PillCam™ SB3 System,

    Medtronic, “PillCam™ SB3 System,” https://www.medtronic.com/covidien/en-nz/ products/capsule-endoscopy/pillcam-sb-3-system.html/, 2025, [Online; accessed 7- May-2025]

    work page 2025

  5. [5]

    Adaptive variable-length coding for efficient compression of spacecraft television data,

    R. Rice and J. Plaunt, “Adaptive variable-length coding for efficient compression of spacecraft television data,” IEEE Transactions on Communication Technology , vol. 19, no. 6, pp. 889–897, 1971

    work page 1971

  6. [6]

    Adaptive run-length/golomb-rice encoding of quantized generalized gaussian sources with unknown statistics,

    H. S. Malvar, “Adaptive run-length/golomb-rice encoding of quantized generalized gaussian sources with unknown statistics,” in Data Compression Conference (DCC’06). IEEE, 2006, pp. 23–32

    work page 2006

  7. [7]

    Kvasir-capsule, a video capsule endoscopy dataset,

    P. H. Smedsrud, V . Thambawita, S. A. Hicks, H. Gjestang, O. O. Nedrejord, E. Næss, H. Borgli, D. Jha, T. J. D. Berstad, S. L. Eskeland et al., “Kvasir-capsule, a video capsule endoscopy dataset,” Scientific Data, vol. 8, no. 1, p. 142, 2021

    work page 2021

  8. [8]

    Galar-a large multi-label video capsule endoscopy dataset,

    M. Le Floch, F. Wolf, L. McIntyre, C. Weinert, A. Palm, K. V olk, P. Herzog, S. H. Kirk, J. L. Steinh ¨auser, C. Stopp et al. , “Galar-a large multi-label video capsule endoscopy dataset,” Scientific Data, vol. 12, no. 1, p. 828, 2025

    work page 2025

  9. [9]

    Recent developments in wireless capsule endoscopy imaging: Compression and summarization techniques,

    B. Sushma and P. Aparna, “Recent developments in wireless capsule endoscopy imaging: Compression and summarization techniques,” Computers in Biology and Medicine, vol. 149, p. 106087, 2022

    work page 2022

  10. [10]

    An improved yef-dct based compression algorithm for video capsule endoscopy,

    A. Mostafa, T. Khan, and K. Wahid, “An improved yef-dct based compression algorithm for video capsule endoscopy,” in 2014 36th Annual International Con- ference of the IEEE Engineering in Medicine and Biology Society . IEEE, 2014, pp. 2452–2455

    work page 2014

  11. [11]

    An ultra-low-power image compressor for capsule endoscope,

    M.-C. Lin, L.-R. Dung, and P.-K. Weng, “An ultra-low-power image compressor for capsule endoscope,” Biomedical engineering online , vol. 5, no. 1, p. 14, 2006

    work page 2006

  12. [12]

    Lossless and low-power image compressor for wireless capsule endoscopy,

    T. H. Khan and K. A. Wahid, “Lossless and low-power image compressor for wireless capsule endoscopy,” VLSI design, vol. 2011, no. 1, p. 343787, 2011

    work page 2011

  13. [13]

    Hardware-efficient low-power image processing system for wireless capsule endoscopy,

    P. Turcza and M. Duplaga, “Hardware-efficient low-power image processing system for wireless capsule endoscopy,” IEEE journal of biomedical and health informatics, vol. 17, no. 6, pp. 1046–1056, 2013

    work page 2013

  14. [14]

    Energy-efficient image compression for capsule endoscopy using a cnn-based feature learning algorithm,

    S. Harshitha, U. Mahadevaswamy, and M. Srikantaswamy, “Energy-efficient image compression for capsule endoscopy using a cnn-based feature learning algorithm,” Engineering, Technology & Applied Science Research , vol. 15, no. 5, pp. 26 217– 26 223, 2025

    work page 2025

  15. [15]

    Automatic detection of informative frames from wireless capsule endoscopy images,

    M. K. Bashar, T. Kitasaka, Y . Suenaga, Y . Mekada, and K. Mori, “Automatic detection of informative frames from wireless capsule endoscopy images,” Medical Image Analysis, vol. 14, no. 3, pp. 449–470, 2010

    work page 2010

  16. [16]

    Categorization and segmentation of intestinal content frames for wireless capsule endoscopy,

    S. Segui, M. Drozdzal, F. Vilarino, C. Malagelada, F. Azpiroz, P. Radeva, and J. Vitria, “Categorization and segmentation of intestinal content frames for wireless capsule endoscopy,”IEEE Transactions on Information Technology in Biomedicine, vol. 16, no. 6, pp. 1341–1352, 2012

    work page 2012

  17. [17]

    Identification of circular patterns in capsule endoscopy bubble frames,

    H. Mir, V . Sadeghi, A. Vard, and A. M. Dehnavi, “Identification of circular patterns in capsule endoscopy bubble frames,” Journal of Medical Signals & Sensors , vol. 14, no. 5, p. 15, 2024

    work page 2024

  18. [18]

    Segmentation and region quantification of bubbles in small bowel capsule endoscopy images using wavelet transform,

    V . Sadeghi, A. Vard, M. Sharifi, H. Mir, and A. Mehridehnavi, “Segmentation and region quantification of bubbles in small bowel capsule endoscopy images using wavelet transform,” Informatics in Medicine Unlocked , vol. 42, p. 101364, 2023

    work page 2023

  19. [19]

    Method and means for recognizing complex patterns,

    P. V . Hough, “Method and means for recognizing complex patterns,” Dec. 18 1962, uS Patent 3,069,654

    work page 1962

  20. [20]

    Edge artificial intelligence wireless video capsule endoscopy,

    A. Sahafi, Y . Wang, C. Rasmussen, P. Bollen, G. Baatrup, V . Blanes-Vidal, J. Herp, and E. Nadimi, “Edge artificial intelligence wireless video capsule endoscopy,” Scientific reports, vol. 12, no. 1, p. 13723, 2022

    work page 2022

  21. [21]

    Smart video cap- sule endoscopy: Raw image-based localization for enhanced gi tract investigation,

    O. Bause, J. Werner, P. Palomero Bernardo, and O. Bringmann, “Smart video cap- sule endoscopy: Raw image-based localization for enhanced gi tract investigation,” in Neural Information Processing . Singapore: Springer Nature Singapore, 2026, pp. 33–47

    work page 2026

  22. [22]

    capsule colonoscopy—a concise clinical overview of current status,

    D. E. Yung, E. Rondonotti, and A. Koulaouzidis, “capsule colonoscopy—a concise clinical overview of current status,”Annals of translational medicine, vol. 4, no. 20, p. 398, 2016

    work page 2016

  23. [23]

    A scalable risc-v hardware platform for intelligent sensor processing,

    P. P. Bernardo, P. Schmid, O. Bringmann, M. Iftekhar, B. Sadiye, W. Mueller, A. Koch, E. Jentzsch, A. Sauer, I. Feldner et al. , “A scalable risc-v hardware platform for intelligent sensor processing,” in 2024 Design, Automation & Test in Europe Conference & Exhibition (DATE) . IEEE, 2024, pp. 1–5

    work page 2024