That’s right. One upside of our use of OpenVINO is Intuiface users can choose to incorporate additional pre-trained models. However, you’d need development skills to 1) incorporate these additional models in the detection server, 2) understand what information these models output, and 3) build interface assets that can work with this information. All 100% doable without our help but unavoidably dependent on coding experience.
Oh, and keep in mind that what we’re doing with OpenVINO is a two-step process. The first step is object detection, the second step is object recognition. For example, today we use a face detection model for the first step and various recognition models (age range, gender, emotion, head pose) for the second step. If you change the model used in the first step - e.g. from face detection to person detection - then you constrain your recognition options - e.g. you’d lose the ability to recognize age/gender.
It is possible to combine multiple detection models in the same server instance (like both face and person), you’re just going to greatly increase the CPU/GPU load. In this case, you’re gonna need one heck of a device to run everything.