Open jcrossyale opened 3 years ago
Or alternatively, is there a way to use OpenFace to conduct sequenced image analysis on a continuously-updating directory?
Not out of the box, you could probably extend the sequence loader with this functionality (e.g. adding monitoring of the directory).
Ideally though, you could try plugging OpenFace directly into whatever is generating the data or piping the data to OpenFace through some form of IPC, but this would also require some work.
Hi there! Hope you are doing well—
I am working on a project that will use OpenFace to do online landmark detection + gaze prediction, however because of certain constraints I am unable to use the built-in online framework (e.g. -device flag for a webcam). The sequence of images (recorded from live video) will be created in real-time.
Are there any tricks to getting the cross-frame normalization functionality from FeatureExtraction with FaceLandmarkImg, using some information about previous frames collected? Or alternatively, is there a way to use OpenFace to conduct sequenced image analysis on a continuously-updating directory?