Banuba has rolled out key enhancements to its Face AR software development kit, targeting improvements in virtual background processing and the addition of face shape analysis. These changes aim to refine user experiences in applications spanning video conferencing, social platforms, and online retail. Developers integrating the SDK now gain tools for smoother visual effects and more tailored AR content.
Enhanced Virtual Background Segmentation
The updated virtual background functionality in Banuba’s Face AR SDK prioritizes edge precision and stability during real-time processing. Traditional segmentation often struggles with motion artifacts like jitter or pixel bleeding, particularly when user clothing matches background hues. Banuba addresses this through advanced neural networks that deliver cleaner foreground isolation, reducing distractions in professional or casual video scenarios.
This refinement supports seamless integration into diverse digital settings without requiring green screens or extensive post-production. The technology processes both portrait and landscape orientations on device, maintaining performance across iOS 13+, Android 6.0+, and web environments with WebGL 2.0 support. Applications include video calls where privacy matters, live streaming for polished broadcasts, and e-learning platforms needing immersive 3D environments.
Technical Underpinnings of Segmentation
Banuba’s approach relies on proprietary deep learning models trained to differentiate human forms from surroundings with high fidelity. Pixels are classified in real time, enabling effects like bokeh blurring or full replacements with images, videos, or GIFs. This on-device computation ensures GDPR compliance by avoiding server uploads, a critical factor for enterprise deployments in regulated sectors.
Performance benchmarks show up to 10x speed gains on Apple hardware compared to prior iterations, alongside compatibility with other SDK modules like face tracking and beautification. For game developers, this means enriched camera feeds during multiplayer sessions; for e-commerce, it facilitates virtual try-ons without visual glitches.
New Face Shape Detection Module
Alongside background upgrades, Banuba introduced a dedicated module for detecting facial contours and proportions. This capability extracts precise measurements of features such as jawline, cheekbones, and forehead width, feeding data directly into downstream AI systems. Developers can access these metrics to customize AR overlays, moving beyond generic filters to user-specific adaptations.
In practice, the module integrates with Banuba’s recommendation engine, which suggests effects based on anatomical data. For instance, it might propose accessories that align naturally with oval versus square face shapes, enhancing realism in virtual makeup or eyewear trials. This personalization boosts relevance in social apps and retail demos, where mismatched effects can deter engagement.
Integration with AI Recommendations
The face shape data transforms recommendation logic from rule-based to deeply contextual. Banuba’s engine cross-references contours with a library of AR assets, prioritizing options that complement individual morphology. This not only improves aesthetic outcomes but also streamlines development workflows, as APIs expose raw detection outputs for custom logic.
Early SDK changelogs highlight synergies with features like skin segmentation and morphing, suggesting face shape detection builds on established 3D modeling. Banuba’s Face Kernel employs direct 3D reconstruction, bypassing 2D-to-3D conversions for superior accuracy under varying lighting and poses.
Broader Implications for Developers
These SDK updates position Banuba as a robust option for cross-platform AR deployment. With support for Unity, Flutter, React Native, and desktop OpenGL 4.3+, enterprises can embed capabilities into hybrid apps without platform-specific rework. Video editors benefit from post-processing modes, while real-time demands in conferencing apps leverage 30 FPS minimums on standard 720p cameras.
Banuba’s nine-year track record includes partnerships with Meta, Gucci, and Samsung, underscoring reliability for high-stakes use cases. The offline processing model appeals to cybersecurity-focused firms, minimizing data exposure risks. Metrics from prior releases, such as nail and teeth segmentation in v1.12.0, indicate a pattern of iterative refinement driving retention.
Use Cases Across Industries
Video conferencing platforms gain professional polish, with clean cutouts elevating remote work aesthetics. Social media apps can deploy hyper-personalized filters, increasing time spent per session. E-commerce sees uplift in conversion via accurate virtual try-ons, where face shape data ensures product visualizations fit realistically.
Gaming and live streaming extend to body-background separation, enabling dynamic environments. E-learning integrates gamified elements, blending AR with educational content for better retention. Photo-video editors access automatic removal tools, democratizing pro-grade outputs for creators.
Company Context and Leadership Perspective
Headquartered in Dubai, Banuba employs around 100 specialists under CEO Dmitry Ogievich, focusing on AR SDKs for face, hair, and hand tracking. The firm’s patented computer vision tech emphasizes device optimization and rapid iteration, as seen in recent lip sync and virtual try-on advancements.
Anton Liskevich, CPO and co-founder, emphasized the dual focus: “Our goal is to make AR interactions flawlessly intuitive and personally engaging. By combining our refined virtual background, which handles the technical challenge of a clean cutout, with intelligent face shape analysis for hyper-personalization, we are empowering developers to build applications that users genuinely love to use.” This aligns with Banuba’s trajectory toward immersive, user-centric tools.
These enhancements expand the SDK’s footprint in competitive AR markets, where precision segmentation and personalization differentiate offerings. Developers worldwide stand to benefit from elevated engagement metrics, as cleaner visuals and apt recommendations foster prolonged interactions.
About Banuba:
Banuba is an augmented reality company with over 9 years on the market, pioneering artificial intelligence, face tracking, virtual try-on, and virtual background technologies. Banuba provides software development kits (SDKs) for augmented reality and video editing, as well as ready-to-integrate solutions for virtual try-on on face, hair and hand.
