RealTime Mode¶
In the Stage Extensions of RealTime Mode, the Model Trainin gand Inference Library provides 6 inference modules. Each requires a model .zip file to be trained and exported from the AI Toolbox first, then loaded in Real-Time Mode:
| Module | Input | Output | Typical Use Case |
|---|---|---|---|
| Image Classification | Image / Camera | Category label + Confidence | Trash sorting, object recognition |
| Object Detection | Image / Camera | Target position + Label + Count | Traffic sign detection |
| Instance Segmentation | Image / Camera | Pixel-level mask + Label | Portrait segmentation, defect detection |
| Pose Classification | Image / Camera | Body pose label | Sports pose recognition |
| Text Classification | Text string | Category label + Confidence | Review sentiment analysis |
| Speech Classification | Microphone audio | Audio category label | Instrument recognition |