AI Eraser

Why It Exists

Users want to remove unwanted objects from photos without complex editing tools. AI Eraser provides a simple, one-tap solution using AI-powered object removal.

How It Works

AI Eraser leverages HyperOS AI's on-device computer vision models to identify and remove unwanted objects from photos. The feature uses semantic segmentation to understand object boundaries and then inpaints the background to seamlessly fill the removed area. Users can simply select the object to remove, and the AI generates a realistic background replacement.

Everyday Use Cases

• Remove a photobomber from a group photo
• Delete an unwanted power line from a landscape
• Remove a watermark or logo from an image
• Clean up a photo by removing distractions

User Workflow

• Open Gallery app and select a photo
• Tap the AI Eraser button
• Select the object to remove or let AI auto-detect
• AI processes and generates the result
• Save the edited photo

AI Processing Flow

On-device computer vision model for object detection and segmentation, followed by AI inpainting for background generation.

Inputs / Outputs

Inputs:

• Photo pixels
• Object selection or auto-detection
• Background context

Outputs:

• Edited photo with object removed
• Generated background fill

Known Limitations

• Requires HyperOS AI-enabled device
• May struggle with complex backgrounds
• Large files may take longer to process

Unsupported Scenarios

• Does not work on older Xiaomi devices without NPU
• May produce artifacts on very complex scenes

Performance Notes

On-device processing is fast for most objects. Processing time varies by image complexity and size.

Available On

Shows where this feature is available and how its AI processing works on each platform. Availability may vary by device.

Research Status

Confidence and verification reflect how complete documentation is. Fields may show Not assessed or Not yet verified while research is ongoing - this flags gaps, not product deficiencies.

Research Notes

Source: Xiaomi HyperOS AI documentation. Uses on-device AI models for privacy.

Details

Capability Mapping

This feature maps to canonical capability family: Image Generation.

Canonical capability: Image Generation.

• Image Generation - Creates novel imagery or emoji-like generated outputs from prompts.

Grounded in 5 device(s) on Xiaomi HyperAI, HyperOS AI.