We are excited to announce official support for the Katana™ Evaluation Kit (KA10000-EVK), the latest low-power edge AI platform from Synaptics. The Katana™ SoC enables low-power AI/ML applications specifically focusing on computer vision, audio, physical sensors or a combination of any of these. Katana can be used to develop and deploy battery-powered intelligent edge AI devices for a variety of market segments, ranging from smart homes, smart cities, smart buildings, manufacturing and Industry 4.0, and consumer electronics.
The platform achieves high power efficiency using its onboard Katana™ low-power Edge AI SoC. The evaluation kit is provided with a separate Himax HM01B0 QVGA monochrome camera module and two onboard zero power Vesper microphones on the left and right side of the board. The board has an embedded STLIS2Dw12 accelerometer and an optional TI OPT3001 ambient light sensor. The connectivity to the board is provided with an IEEE 802.11n ultra-low-power Wi-Fi/Bluetooth 5 combo module. Additionally, there are four Peripheral Module (PMOD) connectors to provide I2C, UART, GPIO, and I2S/SPI interfaces.
The Edge Impulse Studio currently supports the camera, with future support of the microphone.
By supporting the Katana EVK, Edge Impulse opens up new fields of application for low-power edge AI devices allowing users to create their own custom machine learning solutions, from collecting data, training and optimizing AI models and developing applications using the Katana compiler library that enables running inference on the device.
Read the How do I get started? sectionHow do I get started?
To start building your machine learning application using the Katana platform, please consult our getting started guide for the Katana EVK on our documentation page.
You can start by purchasing the Katana evaluation kit, by contacting Synaptics sales. Choose Low-Power AI – Katana Edge AI SoC in the Product drop-down menu.
Once you procure the Katana EVK, connect the platform to the Edge Impulse Studio. Make sure to select the Synaptics KA10000 to obtain the latency calculation for this board:
We have several tutorials for vision and audio, please have a look at the following:
The Synaptics Katana board does support the FOMO object detection model that we have developed at Edge Impulse. To build your FOMO model, take a look at the following tutorial:
Once you have trained your machine learning model, go to the Deployment tab of your Edge Impulse project, then build and download a ready-to-go binary that includes your machine learning model for the Synaptics KA10000; or deploy as a Tensai Flow library and integrate the model into your own firmware!
Congratulations! You have now successfully built and deployed a low-power, industrial-grade embedded machine learning application for the Synaptics Katana EVK.
We are excited to see what you build with the Synaptics Katana platform and Edge Impulse.Post any questions you have and any projects you create over on our forum or tag @EdgeImpulse and @Synaptics on our social media channels!