Computer Vision is a fast growing technology being deployed in nearly every industry. The technology is also driving the evolution of machine learning, edge devices, and human interactions with intelligent systems. As a result, it is becoming critical for computer vision solutions to conduct inference. By conducting inference, computer vision solutions can learn to interpret what the technology records, captures, or observes.
To draw inferences, a system needs to observe measured data. For example, when a person observes a face they can infer certain aspects of identity such as gender or age. Additionally, people can determine attitude and meaning by the angle of a person’s head. In order to identify the same features, angles, and attitudes, computer vision solutions must generate models, learning algorithms, and inference algorithms— and these solutions must interpret input for inference to occur.
The Intel® Distribution of OpenVINO™ toolkit uses deep learning, computer vision, and hardware acceleration and comes with a variety of pre-trained models. The open-source sample videos included below can be used to test these existing models. Models include flaw detection, human detection, vehicle detection, and more. Using the pre-trained models supplied with the Intel® Distribution of OpenVINO™ toolkit, reference implementations, and these sample videos, IoT developers can fast-track their time to production.
Check out the list below and make sure to keep an eye on GitHub for future open-source sample videos
| Video | Description |
|---|---|
| Bolt Detection | This video includes several bolts on a production line conveyor belt. |
| Bolt Multi-size Detection | This video features bolts of different sizes on a production line conveyor belt. |
| Bottle Detection | This video displays 3 bottles removed and replaced at various times throughout the video. |
| Car Detection | This video displays several cars passing through the parking area at different times and angles. |
| Classroom | This video displays students in a classroom in various poses, such as standing, sitting, hand-raising, etc. |
| Face Demographics: Walking and Pause | This video shows different people entering an area, pausing, and exiting in different directions. |
| Face Demographics: Walking | This video shows different people entering an area and exiting in different directions. |
| Fruit and Vegetable Detection | This video displays different fruits and vegetables on a conveyor belt. |
| Head Pose Face Detection: Female and Male | This video displays one female and one male with varying poses during the video. |
| Head Pose Face Detection: Female | This video displays one female face with varying poses during the video. |
| Head Pose Face Detection: Male | This video displays one male face with varying poses during the video. |
| One by One Person Detection | This video displays one person at a time entering a room. |
| People Detection | This video displays people entering an area. |
| Person, Bicycle and Car Detection | This video displays individual people, people with bicycles and cars. |
| Store Aisle Detection | This video displays people examining items in a store aisle. |
| Worker Zone Detection | This video shows workers passing in and out of a defined zone. |
GitHub Reference Implementations Intel® Distribution of OpenVINO™ toolkit
These videos are licensed under Creative Commons Attribution 4.0 International License.