Your Computer Vision Apps...Now Faster
Develop applications and solutions that emulate human vision with the Intel® Distribution of OpenVINO™ toolkit. Based on convolutional neural networks (CNN), the toolkit extends workloads across Intel® hardware (including accelerators) and maximizes performance.
- Enables CNN-based deep learning inference at the edge
- Supports heterogeneous execution across computer vision accelerators—CPU, GPU, Intel® Movidius™ Neural Compute Stick, and FPGA—using a common API
- Speeds up time to market via a library of functions and preoptimized kernels
- Includes optimized calls for OpenCV and OpenVX*
Discover the Capabilities
Deep Learning for Computer Vision
Accelerate and deploy CNNs on Intel® platforms with the Deep Learning Deployment Toolkit (DLDT) that's available in the Intel Distribution of OpenVINO toolkit.
Traditional Computer Vision
Develop and optimize classic computer vision applications built with the OpenCV library or OpenVX API.
Hardware Acceleration
Harness the performance of Intel®-based accelerators: CPUs, GPUs, FPGAs, VPUs, and IPUs.
Who Needs This Product
Software developers and data scientists who:
- Work on computer vision, neural network inference, and deep learning deployment capabilities
- Want to accelerate their solutions across multiple platforms, including CPU, GPU, VPU, and FPGA
Medical Imaging Powered by AI
Intel teamed up with Philips to deliver high performance, efficient deep-learning inference on X-rays and computed tomography (CT) scans without the need for accelerators. The solution runs on servers powered by Intel® Xeon® Scalable processors and was optimized by Intel® Distribution of OpenVINO™ toolkit.
What's New in the 2019 R1 Release
- Supports second generation Intel® Xeon® processors and provides performance acceleration for inference with Intel® AVX512-Deep Learning Boost.
- Extends support to macOS* on the CPU for key toolkit components (Model Optimizer, inference engine, OpenCV, and more).
- Switches parallelism schemes from OpenMP* to Threading Building Blocks (TBB) to provide increased performance in a multinetwork scenario. The most common deployment pipelines run multiple network combinations, and TBB delivers optimal performance for these use cases. For more details, see Intel® Threading Building Blocks as a Default Parallelization.
- Adds support for many new operations in ONNX*, TensorFlow*, and Apache MXNet* frameworks. Provides optimized inference on topologies like Tiny YOLO* version 3, the full DeepLabs* version 3, and bidirectional long short-term memory (LSTM) using the Deep Learning Deployment Toolkit.
- Includes eight pretrained models for new use cases: gaze estimation, action recognition encoder, action recognition decoder, text recognition, and instance segmentation networks.
- Introduces support for binary weights to further boost performance and adds four binary models: ResNet 50, and face, person, and vehicle detection.
- Updates the FPGA plugin to the Deep Learning Accelerator 2019 R1 with new bitstreams for the Intel® Vision Accelerator Design, Intel® Arria® 10 FPGA developer kit, and Intel® Programmable Acceleration Card with Intel® Arria® 10 GX FPGA. Supports automatic scheduling between multiple FPGA devices.
What You Can Do
License Plate Recognition Demonstration
This video discusses high-level features of the toolkit and demonstrates how to run vehicle detection and license plate recognition.
People Counter Reference Implementation
Create a smart video application using models and inference to run single-class object detection.
Interactive Face Recognition
This example demonstrates face detection capabilities, featuring the ability to determine age, gender, and the position of a head.
Awarded by the
Embedded Vision Alliance*
Support
The developer community and our tech experts provide the help you need in this public forum.
Open-Source Software
The OpenVINO™ toolkit is an open-source product. It contains the Deep Learning Deployment Toolkit (DLDT) for Intel® processors (for CPUs), Intel® Processor Graphics (for GPUs), and heterogeneous support. It includes an open model zoo with pretrained models, samples, and demos.
For more complete information about compiler optimizations, see our Optimization Notice.
