Featureless Object Scan

What is Featureless Object Scan?

Featureless Object Scan is a great alternative to Photo Scan Mode as it possesses the ability to reconstruct objects of shiny/reflective surfaces into detailed 3D models, which is extremely challenging for traditional photogrammetry.

Featureless Object Scan is a unique scanning method of the KIRI Engine. Powered by its Neural Surface Reconstruction (NSR) technology, KIRI Engine is able to scan objects of shiny nature into delicate 3D models. It leverages a neural network to anticipate the appearance of an object from various viewpoints, which then reconstructs this depiction into a surface mesh. This scanning method excels in generating detailed 3D models of shiny or reflective objects with fewer feature points, increasing the versatility in the 3D scanning creation pipeline.

What is Neural Surface Reconstruction (NSR) Technology?

NSR technology analyzes sequential images (video) captured from different angles around an object. Unlike traditional photogrammetry, which relies on identifying common features in overlapping photos, NSR employs a neural network to create a volumetric representation of the object, which is then reconstructed into a 3D model. This AI-driven approach allows for precise scanning of objects with minimal features, providing a significant advantage over conventional photogrammetry.

Featureless Object and its 'Featureless Object Scan'

Why use Featureless Object Scan?

KIRI Engine's exclusive Featureless Object Scan offers unparalleled capabilities for scanning objects that are challenging for traditional photogrammetry, especially for objects with shiny/reflective/transparent surfaces.

This includes:

• Shiny and Reflective Surfaces: Objects like aluminum cans, plastic bottles, and toy cars are often challenging for photogrammetry and will result in inadequate depictions due to their reflective nature. Featureless Object Scan excels in capturing these surfaces accurately.
• Simple Geometries: Items with smooth, featureless surfaces, such as ceramics or glassware, benefit greatly from this mode, as the NSR technology can reconstruct these shapes without relying on distinct visual features.
• Transparent Objects: Traditional photogrammetry struggles with transparency. Featureless Object Scan can handle these "simple" complexities, because it uses machine learning to overlook the tranparency and anticipates the object's features.

Featureful vs. Featureless Objects

How to Use Featureless Object Scan?

Preparing the Object for Scanning

Ensure the object is clean and free from obstructions. Proper lighting is essential to avoid excessive shadows. It's best practice to place the object on an elevated surface, where all sides of the object can be effectively captured.

Capturing Videos

Record a video slowly orbiting around the object, covering top, middle, and bottom layers. This comprehensive coverage is crucial for the NSR algorithm to reconstruct the object accurately, as it will use the continuous framing to reconstruct the surface mesh.

Processing and Exporting

Upload the captured video via the KIRI Engine app. Its cloud-based algorithm will analyze the footage and generate a detailed 3D model. Users can fine-tune and edit the model within the app before exporting it in various formats: OBJ, FBX, STL, USDZ, GLTF, GLB, PLY, XYZ.

KIRI Engine's Featureless Object Scan offers a groundbreaking solution for scanning objects with challenging surface characteristics. Tackling traditionally challenging surfaces/objects in the realm of 3D scanning, KIRI Engine's Featureless Object Scan expands the mesh creation possibilities for 3D creators and enthusiasts.