On Tue, Aug 2, 2022 at 1:16 PM Dino Farinacci <[email protected]> wrote:
> Dear Professors Darrell and Yu, > > ... > > Detections from each area are aggregated in algorithmically (location) > addressable shards. > > > > A consolidation process is applied to merge multiple detections from > multiple points of view, varying time-stamps, and varying detection and > localization errors. The consolidation process emerges the current state - > enumeration of the condition of each grid tile aggregated by the shard. > Both condition enumeration, data-clustering, and consolidation processing > applied on network edge computers are aligned with BDD research. > > One question, what if two detections, roughly at the same time, report > different visualizations? Is there a policy, such that if one detected > nothing and another detected an object, that you err on choosing there was > an object present? > > In general such policies are called "non-maximal suppression", and yes, it is standard to include various heuristics and/or learned decision fusion techniques to resolve unique detections. While this is still somewhat an area of research to perfect such techniques, they have been already widely deployed for over a decade in e.g., vehicle and pedestrian detectors, both within views and across multiple views. Each vendor will likely implement a specific policy at first, these policies will generally rely on the network delivering as many samples as possible with minimal disruption to the same AI context, ie EID in this case. > ... > > Therefore we believe that [email protected] is the appropriate review venue > for this draft. Please do not hesitate to contact us for further > discussion of this important topic. > > That is great news. We will make sure we contact you if we need any > questions answered about the use-case. But Sharon is very fluent with the > use-case so he answers most of our questions. > > Cheers and thanks again, > Dino > > Thanks/Cheers t
_______________________________________________ lisp mailing list [email protected] https://www.ietf.org/mailman/listinfo/lisp
