Published February 20, 2026 | Version v1
Preprint Open

Bit-Parallel Morphological Edge Detection via 1024-bit Virtual Registers

Description

14.8-megapixel +617 fps OpenCv x 63X

Based on a 1024-bit virtual register architecture (uint1024_virtual) implemented in standard C++. The design packs 64 binary pixels into each uint64_t word and groups 16 words into a single 128-byte register, allowing the entire working unit to reside within L1/L2 cache. Horizontal edge detection is reduced to a bitwise between a row register and its one-bit left shift, resulting in a single-pass O(W/64) operation that requires no floating-point arithmetic.

On a 14.8-megapixel image (4704 × 3136 pixels), the pure CPU kernel achieves a latency of 1.62 ms (≈617 FPS), outperforming OpenCV Canny by 63×, OpenCV Sobel by 12.7×, and equivalent NumPy XOR implementations by 2.1×. The method has been validated on real photographic scenes, including high-detail natural environments and structured human subjects, producing visually complete and coherent edge maps.

The uint1024_virtual register is a general-purpose computational primitive protected as proprietary intellectual property. This publication establishes prior art for its application to image processing. Broader applicability of the architecture to additional computational domains is reserved for future work.

Potential applications include instantaneous AI pre-processing on edge devices, real-time computer vision for robotics and drones, and high-resolution on-device image analysis without GPU acceleration or cloud dependency.

STATEMENT OF PRIOR ART AND LICENSE TERMS

1. Academic Use and Research Freedom

The author expressly encourages academic and scientific use of this work. Permitted non-commercial activities include research, benchmarking, independent validation, publication of analyses or comparisons, educational use, inclusion in academic pipelines, and theoretical or experimental extensions. All academic use must include appropriate citation to the original work.

2. Statement of Prior Art

This document constitutes a public disclosure of the bit-parallel image processing methodology and the uint1024_virtual architectural approach. The concepts are disclosed solely to establish Prior Art and prevent third-party patent claims under 35 U.S.C. § 102 and international equivalents.

3. Software License

While the conceptual methods are disclosed defensively, the source code, implementations, binaries, and hardware realizations are not in the public domain and are licensed under the PolyForm Noncommercial License 1.0.0.

Permitted (Non-Commercial)
• Academic research and experimentation
• Peer review and independent verification
• Educational and non-profit use
• Non-commercial open-source research implementations

Condition: Any resulting publication must cite the canonical DOI or primary reference.

Prohibited (Commercial)
• Integration into proprietary software or hardware
• Deployment in commercial products, services, or platforms
• Use in paid tools, AI pipelines, or production systems
• Sublicensing, redistribution, or sale of the code or derivatives

4. Commercial Licensing

Any commercial use requires explicit written authorization from the author.

5. No Code-Size Threshold (No De Minimis)

The PolyForm Noncommercial License imposes no exemptions based on code length, fragment size, or proportion of reuse. Any use—partial or complete—remains fully subject to the license.

6. Anti-Snippet Laundering and Anti-Circumvention

Extraction, paraphrasing, refactoring, translation, or functional reimplementation of any component—including bit-parallel structures, register architecture, or execution logic—shall be considered derivative use. Attempts to evade restrictions through minimal reuse, language changes, modular embedding, or functional equivalence do not limit the applicability of the license.

7. Presumption of Derivation

Any system exhibiting substantial functional, structural, or performance similarity, developed after exposure to this work, shall be presumed derivative. The burden of proof for independent creation rests entirely on the alleged infringing party.

8. Knowledge Contamination

Exposure to the source code, binaries, documentation, or technical description constitutes knowledge contamination. Subsequent implementations by exposed parties are not considered clean-room unless supported by contemporaneous evidence of prior independent development.

9. Waiver of Jury Trial

To the fullest extent permitted by law, each party waives the right to a jury trial in any dispute arising from or relating to this license or the software.

10. Severability and Survival

If any provision is held unenforceable, the remaining provisions remain in full force and effect. The following provisions survive termination: license scope, noncommercial restrictions, anti-circumvention, presumption of derivation, knowledge contamination, intellectual property ownership, waiver of jury trial, and remedies.

 

Files

PiroloBitParallel_EdgeDetection_2026.pdf

Files (45.2 MB)

Name Size Download all
md5:13523004f6e5eacc73049529832cb2dc
44.0 MB Download
md5:ba0bebeb4658854795014c3743b0399b
455.7 kB Preview Download
md5:c3d91eed9c02b4bd91ff0d2b7a591f9a
766.9 kB Preview Download

Additional details

Software

Repository URL
https://zenodo.org/records/18706964
Programming language
C++ , Julia , Python
Development Status
Active