Ignitarium Releases Pre-trained AI Applications Library for Renesas RZ/V2L
Ignitarium is excited to release sample vision AI applications targeted for Renesas RZ/V2L SMARC Dev Kit. We are a proud partner of Renesas RZ Ready Partner Network, focusing on developing highly optimized 2D & 3D sensor-based AI applications.
Renesas RZ/V2L is an industrial-grade, energy-efficient Linux micro-processor that is designed for vision AI applications, leveraging powerful hardware acceleration through its Dynamically Reconfigurable Processor (DRP) and multiply-accumulate unit (AI-MAC).
It features a dual-core Arm Cortex-A55 processor, which runs at 1.2 GHz and Arm Mali-G31 GPU, in addition to the DRP-AI IP.
RZ/V2L has a hardware accelerator for AI inference and image processing functions that enable color correction, noise reduction, etc. This helps users to implement vision-based AI applications without any need of an external image signal processor (ISP).
Our team at Ignitarium have been working on a set of sample applications which, in turn, provides developers a quick start for implementing different types of vision AI applications and for exploring its hardware accelerator features. Some of the applications developed have been hosted in the Github library here: https://github.com/Ignitarium-Renesas/RZV2L_AiLibrary
The target evaluation kit (EVK) consists mainly of RZ/V2L board and Google Coral camera module. The rest of the peripherals like HDMI monitor, ethernet cables, USB keyboard and mouse can be connected with the board easily. Further details on getting started on the board can be found here: https://github.com/Ignitarium-Renesas/RZV2L_AiLibrary#readme
The AI library package currently lists the following six applications. More applications will be added to the repository in the coming months.
Human Head Counter
This application shows how a deep learning algorithm is used for simple object detection use case. These applications prove useful in indoor spaces like lifts, escalators and malls to estimate occupancy.
Line Crossing Object Counter
This application highlights the need of good post processing algorithm like object tracking in a sample video analytics pipeline.
Elderly People Fall Detection
The application is built around a deep learning model that detects human key points like head, eyes, shoulder, etc. By analysing positions of these key points, it is possible to detect whether a person is standing or has fallen on the ground. An alert can be raised in case a fall is detected. Such kinds of applications can be improved further and can find place in elderly care homes and hospital treating elderly patients.
Safety Helmet and Vest Detection
This sample application shows the usability of vision AI systems at construction sites. A deep learning-based object detection neural network is trained to detect safety helmets and safety vests.
Human Age and Gender Detection
This sample application shows the application of a deep neural network that can classify a closeup facial image of a person into any age group. It can also classify the facial image into an obvious looking male class or an obvious looking female class.
Face Recognition and Spoof Detection
The application shows the capability of a deep neural network to identify whether the human present in front of the camera is a real human or just a photo of the person. If it is real human, then application goes on to check the pre-registered faces and recognizes the person.
All these sample applications are developed using open-source deep learning models and open-source training data. There are detailed instructions that allow developers to quickly evaluate the existing AI applications on RZ/V2L board as well as to experiment with other types of applications using these Deep Learning models. Due to the inherent limitation of small datasets and state of DL models, the accuracy of the applications may be low. If there is a need to develop a productized version of any of these or other vision AI applications on RZ/V series of AI SoCs, please reach out to the teams at Ignitarium or Renesas, and we will enable the same.
We are excited to see what interesting vision AI applications you would develop on RZ/V2L. Please post your queries using Issues or Discussions sections on Github or you can tag @ignitarium or @RenesasGlobal on social media. We would also like to hear about new DL models and applications that you want to see supported in our roadmap.
Stay tuned for more additions to the github page.
Ready Resources:
RZ/V2L Overview and Documentation
Ignitarium-Renesas RZ/V2L github
