search

mikel-brostrom / boxmot

BoxMOT: pluggable SOTA tracking modules for segmentation, object detection and pose estimation models
GNU Affero General Public License v3.0
6.67k stars 1.71k forks source link

How to incorporate a black box Reid model ? #1311

Closed ozayr closed 7 months ago

ozayr commented 7 months ago

Search before asking

Question

Hi,

I am wondering if one could simply add a ReID model that you know output the features for a given input eg.

I want to use the face Reid model from this repo https://github.com/PINTO0309/PINTO_model_zoo/tree/main/432_face-reidentification-retail-0095 how could I simple drop in the Reid model (it is in onnx format) ? The model outputs a Nx256 feature vector where N is the input images

currently im limited to the ReID models hardcoded into the repo.

could there be some kind hook that we could use where we could implement the feature extractor portion of the tracker

mikel-brostrom commented 7 months ago

You need to know the preprocessing required for the model's input as well. It is not possible to just plug the model into the repo. Unless the preprocessing is embedded in the model, which is usually not the case

github-actions[bot] commented 7 months ago

👋 Hello, this issue has been automatically marked as stale because it has not had recent activity. Please note it will be closed if no further activity occurs. Feel free to inform us of any other issues you discover or feature requests that come to mind in the future. Pull Requests (PRs) are also always welcomed!