Aiarty Image Enhancer leverages Core ML for fast and private AI enhancement. Balances workloads across CPU, GPU, and ANE to ensure the best performance on your Mac.
As the demand for high-quality image enhancements across photography, AI-generated content, and digital restoration continues to grow, macOS users are increasingly seeking solutions that balance performance, privacy, and efficiency. Aiarty Image Enhancer, developed by Digiarty Software, addresses this need by bringing advanced on-device AI image processing to the Mac through tight integration with Apple’s Core ML framework.
By moving the entire inference pipeline to Core ML, Aiarty will be able to take advantage of the full potential of Apple’s hardware, especially the Neural Engine. ”
— Jack Han, Aiarty CEO
Powered by Core ML and optimized to take advantage of the CPU, GPU, and Apple Neural Engine (ANE), the software delivers fast performance while ensuring complete user privacy with complete on-device processing.
Leverage Apple Core ML for on-device intelligence
Apple’s Core ML is a machine learning framework optimized for on-device performance that enables developers to efficiently run AI models on macOS and iOS. Aiarty Image Enhancer leverages Core ML to run AI models directly on macOS devices, enabling fast, secure, fully on-device image processing without relying on cloud servers.
Also read: AiThority Interview with Glenn Jocher, Ultralytics Founder and CEO
By intelligently distributing workloads across CPUs, GPUs, and ANEs, the software dynamically adapts to each machine’s hardware capabilities to optimize performance. Introduced to Apple Silicon with the M1 chip, ANE is specifically designed for AI and machine learning tasks, delivering high speed processing with low power consumption.
“Moving the entire inference pipeline to Core ML will allow Aiarty to take full advantage of the full potential of Apple’s hardware, especially the Neural Engine,” said Jack Han, CEO of Aiarty. “Users can now upscale, denoise, sharpen, and restore images all on their own devices without sending files to the cloud and with significantly reduced latency.”
Flexible performance modes across Mac devices
To strike the perfect balance between performance and compatibility, Aiarty offers three different inference modes tailored to different hardware and operating systems.
• All-unit mode: Automatically selects the best compute unit based on device and model needs and runs on two or three units of CPU, GPU, and ANE simultaneously.
• CPU + GPU mode: Recommended mode that balances high performance with wide compatibility.
• CPU-only mode: slower, but ensures wide compatibility, especially on older systems.
Available modes vary depending on the macOS version.
• Systems running macOS 13 or later support all three inference options, including the powerful “All Units” and “CPU+GPU” modes.
• Systems running macOS 12 or lower are strictly limited to “CPU only” mode.
• If your Mac system is older than macOS 13, you won’t be able to take advantage of GPU or ANE acceleration, which will significantly impact inference performance and speed.
This clear line allows users of the latest Apple Silicon and updated macOS versions to fully experience the uncompromised speed of Aiarty’s GPU and ANE optimizations, while older Intel-based Macs and legacy systems maintain functional compatibility.
Designed for real-world image enhancement workflows
Aiarty Image Enhancer provides a comprehensive set of AI-powered tools, including:
• Upscale AI images up to 16K on Mac
• Intelligent noise reduction to maintain natural texture
• Deblur and restore images with low quality or motion blur
• High-fidelity face correction for clearer, more natural-looking portraits
• Built-in color adjustment tools for complete image optimization
• SDR to HDR conversion improves dynamic range, contrast, and visual depth.
• AI-powered object removal for clean, uncluttered visuals
• Adjustable intensity slider for fine-tuning and precise control of reinforcement intensity
• Supports high-volume batch processing for efficient large-scale workflows.
Powered by multiple AI models, including More-Detail GAN, AIGCsmooth, and Real-Photo, the software is designed to process a variety of image types, from real-world photos to AI-generated visuals.
Also read: The infrastructure war behind the AI boom
[To share your insights with us, please write to psen@itechseries.com]
