|
IVALab Python Libraries
Collection of code for computer vision and robotics with specific API.
|
Perception code that aims to extract interpretable signals from a visual stream. More...

Modules | |
| Reports | |
| A set of interfaces for reporting perceiver outcomes during processing. | |
Classes | |
| class | BuildCfgMonitor |
| Build configuration instance for a monitor. More... | |
| class | BuildCfgPerceiver |
| Build configuration instance for a perceiver. More... | |
| class | BuildCfgProgress |
| Build configuration instance for a progress monitor. More... | |
| class | CfgMonitor |
| Configuration instance for a Monitor. More... | |
| class | CfgPerceiver |
| Configuration instance for a perceiver. More... | |
| class | CfgProgress |
| Configuration instance for a Progress monitor. More... | |
| class | Monitor |
| A simple interface class for monitoring the outcomes of a perceived scene. More... | |
| class | Perceiver |
| Basic implementation of a perceiver class. More... | |
| class | Progress |
| A simple interface class for progress monitoring a perceived scene. More... | |
Perception code that aims to extract interpretable signals from a visual stream.
The most basic perceiver consists of a detector and a tracker, and returns some vector quantity related to a detected instance in the image. Adding a track filter will, in principle, provide a cleaner output signal less affected by noise or measurement uncertainty. More complex implementations follow.
Sometimes the tracked signals are not important, but rather their meaning. In that case an activity/action detector or recognizer will convert the tracked signal(s) into semantically meaningful labels or their equivalent. A Monitor packages these two together.
Likewise, there may be a need to compare the perceived state to some target or goal state. The most common reason for this comparison would be to assess completion or proximity to the target/goal state. In that case a Progress monitor is needed.
This package attempts to encapsulate the generic functionality of these different classes (Perceiver, Monitor, Progress monitor).