Image Compression Techniques for Wireless Capsule Endoscopy

Wireless capsule endoscopy (WCE) relies on ingestible imaging devices that traverse the gastrointestinal tract, capturing high-resolution frames for diagnostic purposes. The strict power, size and bandwidth constraints of these capsules necessitate highly efficient image compression strategies. Techniques fall broadly into lossless and (near-)lossy categories. Lossless schemes guarantee full reconstruction of original frames, employing predictive coding, entropy encoding such as Golomb–Rice or Huffman, and colour-space conversions tailored to limited silicon resources. Near-lossless methods accept minimal distortion to achieve higher compression ratios, focusing on region-of-interest preservation to prioritise diagnostic detail. Transform-based codecs exploit spatial and temporal correlations via two-dimensional or three-dimensional discrete cosine transforms, often enhanced by zigzag scanning and adaptive quantisation. Distributed video coding paradigms have also been adapted to offload complexity to external readers. Hardware implementations must balance gate count, memory footprint and energy consumption, driving innovations in multiplier-less architectures, subsampling strategies and raw Bayer CFA processing. Recent advances highlight the interplay between algorithmic ingenuity and low-power VLSI design, enabling real-time wireless transmission of clinically viable image streams.

Research from Nature Portfolio

No recent Nature Portfolio content available.

Research from all publishers

Foundational work introduced a lossless compression pipeline combining raster-scan predictive coding with variable-length Golomb–Rice and unary encoders, implemented on field-programmable gate arrays. This system achieved compression ratios exceeding 75 % without sacrificing image fidelity, validating performance in in vivo trials. The approach eliminated large buffer requirements through a low-complexity colour-space converter optimised for white-light and narrow-band imaging.

Building on transform techniques, a three-dimensional discrete cosine transform method reconstructed Bayer CFA data into volumetric blocks, applying an optimised 4-point DCT butterfly structure to minimise multipliers. Enhanced quantisation, zigzag scanning and frequency-domain filtering suppressed blocking artefacts, delivering average compression ratios near 23:1 and peak signal-to-noise ratios above 40 dB with low computational overhead.

Recent near-lossless designs have prioritised silicon area and energy efficiency by processing raw Bayer data and exploiting inter-component colour correlations in entropy encoding. A silicon-proven intellectual property core implemented adaptive Golomb–Rice coding in a 180 nm CMOS process, reducing memory requirements by almost 90 % compared with JPEG-LS-based designs. The completed system consumed under 25 μJ per 512 × 512 frame while preserving diagnostically relevant detail.

Image Compression Techniques for Wireless Capsule Endoscopy publication trend

The graph below shows the total number of articles in image compression techniques for wireless capsule endoscopy across all publications each year (not limited to Nature Index journals).

Technical terms

Wireless capsule endoscopy: Ingestible medical device capturing and wirelessly transmitting images of the gastrointestinal tract.

Compression ratio: Measure of original data size relative to compressed data size.

Lossless compression: Encoding that allows exact reconstruction of the original image.

Near-lossless compression: Encoding that permits minimal, controlled distortion to enhance compression.

Discrete cosine transform (DCT): Mathematical transform that expresses image data in terms of frequency components.

Entropy encoding: Lossless coding methods that assign shorter codes to frequent symbols (e.g. Golomb–Rice, Huffman).

Bayer Colour Filter Array (CFA): Mosaic arrangement of colour filters on an image sensor capturing raw RGB data.

Peak signal-to-noise ratio (PSNR): Metric quantifying reconstructed image quality relative to the original.

References

1. Design of a Lossless Image Compression System for Video Capsule Endoscopy and Its Performance in In-Vivo Trials. Sensors (2014).
2. 3D DCT Based Image Compression Method for the Medical Endoscopic Application. Sensors (2021).
3. Wyner-Ziv video coding for wireless lightweight multimedia applications. EURASIP Journal on Wireless Communications and Networking (2012).
4. Low-Power Low-Area Near-Lossless Image Compressor for Wireless Capsule Endoscopy. Circuits, Systems, and Signal Processing (2022).