Problem
In deploying and extending DTPR, we have encountered computer vision use cases that are configured to detect objects or changes in conditions, rather than people.
For example, a computer vision system may be used to identify:
- Litter
- Overflowing bins
- Abandoned objects or items left in a space
- Whether a parking space, loading zone, or curb area is occupied
Right now, the closest element is Person Detection, but it is not a semantically accurate fit for these use cases and may mislead readers into thinking the system is detecting people, when its actual function is to identify non-human objects or environmental conditions.
This creates a gap in the taxonomy and makes it harder to describe certain computer vision deployments precisely and consistently.
Proposal
We propose adding Object Detection as a new DTPR element in the Technology category. This would represent computer vision systems used to identify or classify non-human objects, materials, or environmental conditions.
Rationale
Adding this element would improve semantic precision in the taxonomy and reduce potential confusion in public-facing materials.
More broadly, this reflects how the Technology category functions in DTPR. It is intended to help people understand what a system is doing, not only identify the underlying software/hardware. DTPR research has consistently shown that a central public concern is whether a technology can see them or otherwise use them as data. In that context, distinguishing Object Detection from Person Detection makes the standard more legible and more responsive to the questions people have, even if both rely on computer vision.
Problem
In deploying and extending DTPR, we have encountered computer vision use cases that are configured to detect objects or changes in conditions, rather than people.
For example, a computer vision system may be used to identify:
Right now, the closest element is Person Detection, but it is not a semantically accurate fit for these use cases and may mislead readers into thinking the system is detecting people, when its actual function is to identify non-human objects or environmental conditions.
This creates a gap in the taxonomy and makes it harder to describe certain computer vision deployments precisely and consistently.
Proposal
We propose adding Object Detection as a new DTPR element in the Technology category. This would represent computer vision systems used to identify or classify non-human objects, materials, or environmental conditions.
Rationale
Adding this element would improve semantic precision in the taxonomy and reduce potential confusion in public-facing materials.
More broadly, this reflects how the Technology category functions in DTPR. It is intended to help people understand what a system is doing, not only identify the underlying software/hardware. DTPR research has consistently shown that a central public concern is whether a technology can see them or otherwise use them as data. In that context, distinguishing Object Detection from Person Detection makes the standard more legible and more responsive to the questions people have, even if both rely on computer vision.