Information Technology Standards Summary - September 2025

Looking back at September 2025, Information Technology and Office Equipment professionals witnessed the publication of a landmark standard: ISO/IEC 6048-1:2025. As AI-driven solutions continued to redefine data compression and multimedia processing, this sole release highlighted a focused shift toward learning-based, interoperable imaging frameworks. This retrospective offers a deep-dive into the month's activities, the driving themes in standardization, major features of JPEG AI, and strategic insights for organizations seeking to future-proof their image processing pipelines and compliance regimes.

Monthly Overview: September 2025

September 2025 represented a distinctive moment in the Information Technology and Office Equipment sector, with the standards landscape characterized by a focused, high-impact publication rather than sheer volume. The release of ISO/IEC 6048-1:2025 illuminated the growing importance—and maturity—of Artificial Intelligence in image compression technology. Unlike prior months that may have featured a mix of security, data exchange, or software lifecycle standards, this period centered on advancing the JPEG family with a learning-based approach built for dual human and machine vision contexts.

This month's activity signaled a clear industry trend: the integration of deep learning models into core data representation schemes to serve not only traditional viewing needs but also machine-vision pipelines in computer vision and AI analysis tasks. The specificity of the standard, developed collaboratively by ISO/IEC and ITU-T, emphasized high interoperability, the importance of performance at scale (covering diverse applications and bit depths), and the alignment of image compression with emerging AI workloads. For organizations, September 2025's publication underscored the need to reconsider their multimedia processing benchmarks—and illustrated that future-ready standards now require a balance of technical sophistication and flexibility in implementation.

Standards Published This Month

ISO/IEC 6048-1:2025 - JPEG AI Learning-Based Image Coding System — Part 1: Core Coding System

Full Standard Title: Information technology — JPEG AI learning-based image coding system — Part 1: Core coding system

Publication Date: 2025-09-01

Organization: ISO/IEC, in cooperation with ITU-T

Scope & Objectives: ISO/IEC 6048-1:2025 establishes the core technical specification for JPEG AI, an image coding system leveraging deep learning and neural networks to deliver enhanced compression for both human and machine vision applications. The standard targets efficient, single-stream, compressed domain representations, ensuring significant improvements in subjective image quality—especially at lower bitrates—compared to legacy JPEG and other widely adopted image codecs. Its architecture is optimized to enable seamless data interchange across platforms, promote interoperability, and serve diverse modalities, from media broadcast and storage to real-time streaming and computer vision analytics.

The Core Coding System (Part 1) specifies:

• Syntax Format: Defines the structure and semantics of the coded bitstream.
• Decoding Process Requirements: Details how a conformant decoder parses and reconstructs human-viewable images from the compressed stream, including the roles of neural networks.
• Generic Applicability: Designed to accommodate a range of use cases, resolutions, bit depths (8–16 bits), and color sampling formats (4:4:4, 4:2:2, 4:2:0).
• Interoperability: Facilitates image data interchange among varied services and applications, bridging the gap between human visual quality and machine-vision accuracy.
• Profiles and Levels: Introduces profiles and levels for practical implementation, subset compatibility, and performance tuning across the ecosystem.

Key Requirements and Specifications:

• Stream and Decoder Profiles: Provides syntax subsets and decoder capabilities, allowing implementers to tailor deployments based on device requirements and application constraints.
• Single-Stream Architecture: Ensures compressed images contain all necessary information for both display and downstream vision tasks.
• Model Separation: The actual trained neural network parameters required for decoding are available as external attachments rather than embedded—streamlining compliance and enabling updates independent of the core syntax.
• Optional Features: Includes optional codestream segments for future extensibility, such as enhanced rendering or post-processing, without requiring core conformance shifts.
• Algorithm Transparency: Encoding process is not mandated, but compliant decoders must reliably reconstruct the image as specified, with output allowed to deviate only within specified quality margins.

Target Industries and Use Cases:

• Digital imaging, archiving, and storage solutions
• Media distribution and streaming platforms
• Mobile device imaging, including messaging and sharing applications
• Surveillance and visual analytics systems
• Machine vision and autonomous systems
• Cloud-based AI image processing services

Regulatory and Industry Context: JPEG AI (Core Coding System) sits at the intersection of traditional image standards and emerging AI regulations, supporting compliance with global data interoperability mandates while responding to market demands for higher compression efficiency and cross-system portability. It serves as an enabler for cloud-native, privacy-aware, and intelligent image workflows.

Notable Features or Changes:

• First Edition: Marks the inaugural, harmonized release of JPEG AI under ISO/IEC and ITU-T (corresponding to ITU-T T.840.1).
• Deep Learning Backing: Integrates neural networks, attention mechanisms, and context modeling into the decoding pipeline.
• Region of Interest and Progressive Decoding: Supports functional enhancements for modern content delivery and adaptive visualization.
• Extensibility: Optional tools and headers support evolving needs without undermining backward compatibility.

Key highlights:

• Establishes a future-focused, AI-native image coding framework built for interoperability
• Improves compression efficiency—benefiting both storage and transmission cost for human and machine consumption
• Enables compliant decoders to reconstruct images with high fidelity and process readiness for vision tasks

Access the full standard:View ISO/IEC 6048-1:2025 on iTeh Standards

Common Themes and Industry Trends

While September 2025 featured a single release, this standard signaled a pronounced move toward:

• AI in Core Data Representation: The use of trainable neural networks for compression is no longer experimental but fundamental to new standards development. Traditional block-based and handcrafted transforms are now augmented or replaced by model-driven analysis.
• Dual-Purpose Encoding: A unified bitstream for both human visualization and machine vision positions the industry to better support automated analytics, content moderation, and intelligent search without re-encoding.
• Modularity and Extensibility: The separation of model parameters and optional post-processing tools illustrates a trend toward modular standards, enabling agile adaptation as AI models improve or regulatory requirements change.

Sector-specific Impact: The publication suggests renewed focus on applications serving both consumer multimedia (e.g., social media, streaming) and industrial/commercial AI use cases (e.g., surveillance, smart cities, manufacturing QA). Vendors and implementers should anticipate cross-domain deployment expectations and plan for increased demand for machine-readable image compression.

Compliance and Implementation Considerations

For organizations preparing to adopt ISO/IEC 6048-1:2025, the following guidelines apply:

1. Evaluate Decoder Readiness: Ensure that decoders in use, whether in endpoints or back-end services, are capable of parsing the JPEG AI stream and applying the specified neural network models. Where third-party solutions are used, request conformance evidence, especially for interoperability in hybrid environments.
2. Review Implementation Profiles: Select applicable stream and decoder profiles based on organizational needs (e.g., resolution, color format, bit depth) and device capabilities. For applications with strict latency or compute requirements, subset profiles may offer tractable integration paths.
3. Acquire Model Parameters: Obtain and integrate the required neural network weights, hosted externally per the standard’s annexes. Establish a workflow for updating models as they evolve or new editions are released.
4. Plan Integration Timelines: New standards adoption typically entails a multi-phase approach—pilot, phased migration, and full deployment. Proactively engage vendor and supplier partners to align roadmaps with JPEG AI compliance availability.
5. Document Compliance: Build internal documentation referencing the relevant profiles, levels, and use of optional toolsets, ensuring all deployments are audit-ready for procurement or regulatory review.

Resources for Getting Started:

• Access the iTeh Standards portal for the full, official text
• Engage in ISO/IEC JTC 1/SC 29 working group briefings for technical clarification
• Leverage training and interoperability test suites offered by standards consortia

Conclusion: Key Takeaways from September 2025

September 2025’s Information Technology and Office Equipment standardization activity spotlighted a major leap in harmonizing image compression and AI through the publication of ISO/IEC 6048-1:2025. By formalizing a deep-learning-powered, single-stream approach catering to both human and machine needs, the standard establishes a future-proof foundation for digital imaging, media distribution, and computer vision workflows alike.

Recommendations for Industry Professionals:

• Stay informed: Regularly monitor the evolution of learning-based codification and standards supporting AI in imaging.
• Prioritize compliance: Align technology acquisition and development strategies with JPEG AI requirements, especially where interoperability and image analytics are central.
• Invest in readiness: Equip teams with training and resources to implement compliant decoders and leverage the full feature set, including machine-friendly compressed formats.

Keeping pace with standards like ISO/IEC 6048-1:2025 positions professionals to deliver enhanced user experiences, enable smarter machine vision, and drive value across converging technology domains. For further details or to acquire the official specification, visit iTeh Standards.