SMART VISION, EMBEDDED EVERYWHERE.
Devices with embedded systems are all around us today. Our novel face analytics technology architecture allows us the unique advantage of embedding our vision algorithms into embedded systems that power the machines and devices we use everyday.
Emotionally Intelligent Devices.
Our vision towards accelerating the adoption of Ambient Intelligence (AmI) fuels our motivation towards maintaining a perfect balance between continuous innovation and enabling the ease-of-access to our emotion recognition technology.
We strive everyday to accommodate a wide range of applications while maintaining our software’s speed, accuracy and customization. We have a solution for every use case and work with a wide range of sensor and electronic manufacturers to ensure hardware-specific optimizations and cross-compatibility.
We strive to reach a high degree of classification accuracy for each metric we support while maintaining computational efficiency and real-time capabilities especially in embedded environments.
The user has the flexibility to enable or disable any of the core recognition modules or set a compute frequency for each on demand. For instance, the emotion recognition and head pose modules may not be needed on a frame-by-frame basis in certain applications. In these cases, the compute intervals for each module can be set to, say, every 500 milliseconds to reduce the computational burden.
Our services also support processing through a Graphical Processing Unit (GPU), which can increase the processing speed and relieve the computational workload from the CPU. Again, to maximize user flexibility, we provide an option to enable or disable this feature based on the user’s needs. We also expose all parameters that may affect the classification accuracy so that the user can adjust and customize the system performance based on the requirements of the specific application.
Optimized for Perfomance, Power and Area (PPA).
We focus heavily on increasing processing performance with reduced power and area footprint for Embedded System applications. We optimize performance, power and area (PPA) by utilizing all cores – both general purpose cores and specialized cores such as DSPs and GPUs.