AI & Deep Learning

From training pipelines to inference deployment on your systems —
use deep learning directly inside the image processing pipeline.

Inference Engine

C++ inference deployment optimized for your system environment.

An inference engine that runs ONNX models directly in a C++ environment.

Choose the execution backend and accelerate with TensorRT FP16 and engine caching.

GPU memory limits and custom allocator injection to coexist with your system resources.

Output shape prediction for varying input sizes, with batch processing.

Deep learning inference as one stage of the image processing chain.

Insert an inference node inside the image processing pipeline.

Image-to-tensor conversion and batch I/O helpers.

From data preparation to training management.

Data augmentation pipeline construction.

Image normalization and adjustment pipelines for training.

Per-pixel loss weight map support.

Training history, metrics, and statistics management.

A development flow that connects research and product deployment.

Bindings that connect the Python training environment and the C++ deployment environment.

A GUI tool for running and monitoring training.

We will guide you to the right configuration and an evaluation SDK for your system or software.