The DeepView AppPack provides you with the building blocks and glue for robust, turn-key intelligent vision applications, while still allowing you the flexibility to customize and leverage your company’s competitive advantage.
The highly-optimized applications integrate an end-to-end solution, complete with target-specific vision pipeline, fully-optimized neural network models, and DeepViewRT inference engine.
Turn-key applications that combine the key features of ModelPack, VisionPack, and DevPack to enable you with target-optimized end-to-end AI Vision solutions ready for deployment on DeepView Vision Starter Kits or your own hardware built with i.MX Application Processors and i.MX RT Crossover MCUs.
Combining detection, classification, and motion tracking, this application is used to identify and monitor objects belonging to one or more classes, such as people, animals, vehicles, or inanimate objects. Depending on the scene complexity, desired frame rate and available compute, tracking is available on both MCU and high-performance applications processor.
This application uses flexible combinations of detection and segmentation to find and identify visual anomalies. Includes support for adjustment over time to accommodate anomaly acceptance criteria. By retraining the DeepView model with your dataset the application can be quickly adapted for your specific needs. Depending on complexity and desired frame rate, this application can be deployed to MCU's or applications processor.
Occupancy and Utilization
Enhanced detection algorithms verify the number of people, animals or objects in the region of interest. This application can be combined with face recognition to authenticate users and quantify key metrics. Depending on the room size, number of people, and video frame rate, this application can be built on an MCU or applications processor.
This visual intelligence application is available for several domains, including safety, security, surveillance, and efficiency. It can be applied to the workplace or public areas. Due to real-time performance constraints, this application is built on an applications processor.