AI inspection library for embedded devices (DLL and OEM provision)

Added to bookmarks
Bookmarks list

Bookmark has been removed
Bookmarks list

You can't add any more bookmarks

By registering as a member, you can increase the number of bookmarks you can save and organize them with labels.
Free membership registration

DeepAge Sensor

last updated：Aug 25, 2026

Spot

Official site

Built libraries and public C headers. 12ms per frame on CPU only, no GPU required. Over 8 years and more than 100 units deployed.

To manufacturers and system integrators who want to incorporate AI inspection features into devices that handle images: DeepAge Sensor is provided as a pre-built inference library (Windows/UNIX). You only need to reference one public C header (extern "C"). It works simply by linking in C/C++ or loading with ctypes in Python. The CPU version requires no GPU, CUDA, or drivers, and operates at 12ms per image (the GPU version operates at 2ms). Training is done on our cloud, and only the model file is distributed to the devices, so there is no need to rebuild the device software every time the model is updated. An unsupervised anomaly detection library (.uai) can also be used with the same API structure. We will provide a library and sample model for evaluation, so you can first check if create→load→inference works in your environment. OEM/white label offerings are available. Your company will handle primary support, while we will provide secondary support.

Related Link - https://sensor.deepage.net

Inquire About This Product

basic information

【Provided Items】Built inference library (GPU version and CPU version) + public C header (inference.hpp) + model files (.ai/.uai) + complete self-test for operation confirmation 【Operating Requirements】CPU version: No GPU, CUDA, or driver required (12ms per image) / GPU version: NVIDIA sm_86 or later, CUDA 13.2 series (2ms per image) 【OS】Windows / UNIX (Linux x86_64) 【API】Four stages: create → load → inference → close (6 functions). Unsupervised anomaly detection follows the same structure 【Model Update】Only file replacement is needed. No need to rebuild the device software 【Support Responsibility】Primary: Your company / Secondary: Our company (technical separation and bug response)

Price information

OEM royalty (based on the ratio to the equipment's main price) individual quotation / evaluation library and sample model provided.

Price range

Delivery Time

※The initial model creation is free (accuracy report in as little as one week).

Applications/Examples of results

- OEM integration of AI inspection functions into products of inspection equipment manufacturers (over 100 units implemented since the start of provision in 2018, 8 years ago) - Supports both retrofitting inspection options to existing models and standard installation in next-generation models - Proven track record of deployment in equipment for overseas markets, including China

Detailed information

8 years of experience with over 100 devices installed and key specifications.
Three learning methods. The unsupervised anomaly detection library has the same API structure.
Learning is done in the cloud, and inference is performed only by the device's CPU. Model updates are done solely by replacing files.
1. Receipt of evaluation library 2. Operation confirmation with self-test 3. Application integration 4. Model creation in actual work (first time free)
6 industries and various defects addressed (solder defects/scratches/dents/burrs/chips/foreign objects/missing items/printing defects)