Project Aria Documentation Archive

The Project Aria Machine Perceptions Services Command Line Interface (MPS CLI) is a tool used to request and receive Machine Perception Services (MPS). It offers more features and more robust file uploading than the Desktop Companion app. The CLI supports two primary modes:

• Single Mode: Processes each recording individually. Output is saved in the same directory as the input VRS file. This is the most common usage.
• Multi Mode: Processes multiple recordings together to generate multi-sequence SLAM data in a shared coordinate frame. Requires a specified output directory.

Inputs can be individual files or directories, and multiple inputs can be processed in a single command.

Project Aria's Machine Perception Services (MPS) estimate gaze direction using Eye Tracking (ET) camera images. The estimation is performed in the Central Pupil Frame (CPF). Depending on the recording and calibration, MPS produces the following output files:

• summary.json: A high-level report on the eye gaze generation process.
• general_eye_gaze.csv: Gaze data based on the standard eye gaze configuration (available for all recordings with ET cameras enabled).
• personalized_eye_gaze.csv: Gaze data generated specifically for recordings that included in-session Eye Gaze Calibration.

EgoBlur is an open-source AI model from Meta that detects and blurs Personally Identifiable Information (PII), specifically faces and license plates, using FasterRCNN-based detector models.

Available Tools:

• EgoBlur VRS Utilities: A C++ tool for processing Aria VRS files.
• EgoBlur Demo: A Python3 tool for processing PNG, JPEG, or MP4 files.

Emotional Intelligence (EQ) can be categorized by how a person processes information (Knowing vs. Acting) and the target of that focus (Self vs. Others). This creates four distinct quadrants of ability:

• Self-Awareness: Knowing your own emotions.
• Self-Management: Acting on/managing your own emotions.
• Social Awareness: Knowing the emotions and cues of others.
• Relationship Management: Acting on that knowledge to manage interactions with others.

Project Aria glasses are multimodal data collection devices used for spatial AI and AR research. They capture data from a human perspective using the following sensors:

• Five cameras: two Mono Scene, one RGB, and two Eye Tracking cameras.
• Non-visual sensors: two IMUs, magnetometer, barometer, GPS, Wi-Fi beacon, Bluetooth beacon, and microphones.

Data is stored in VRS files, which preserve the integrity of individual sensor streams and ensure they are time-aligned using shared timestamps.

Permaculture principles for sustainable land management can be adapted to documentation and architecture to ensure the process is sustainable and user-centric. Key applications include:

• Listening to the ecosystem: Avoid forcing a specific structure onto documents without considering the specific workplace and user base.
• Obtaining a yield: Focus on smaller, impactful user stories rather than getting lost in overly elaborate structures.
• Using edges and the marginal: Accept that 'messy' is acceptable during the forming and storming phases of a project.
• Energy storage: Prioritize big-picture architectural work over premature copy-editing perfection.

For users who have received Project Aria glasses, the following resources are available for setup and operation:

• Overview: Refer to the 'About ARK' documentation for device usage options.
• Setup: Follow the 'Getting Started With Project Aria Glasses' quickstart guide.
• Manual: Consult the 'Glasses User Manual' for detailed operations, including factory resets.
• Support: Visit the official Support page for troubleshooting and contact options.