On-device AI for battery-powered cameras
Siam Edge Silicon is evaluating person detection as its first on-device AI workload for consumer smart cameras.
Where the chip would fit
The proposed module would process camera frames on the device and report a person-detection event to the camera firmware. The pilot would measure how this affects battery use, detection quality, wake-to-alert time and component cost.
Evaluation architecture
- Image sensor
- Inference on evaluation hardware
- Host MCU or camera firmware
- Wireless radio
What a device partner brings
One target camera model, its current hardware or an agreed baseline, representative operating conditions and acceptance targets. Agree how test footage and device information may be used before sharing them.
What the partner receives
A benchmark comparing both setups under the same conditions, an assessment of integration constraints, and a recommendation on whether a new module or custom chip is worth developing.