search

Streets-Data-Collaborative / groundwork-detection

Use customvision.ai to detect specific elements of urban infrastructure.
2 stars 5 forks source link

Meeting 4 - Notes #15

Open vr00n opened 5 years ago

vr00n commented 5 years ago

Group 1 - Signage

@karansaini282 walked through the Neighborhood benchmarking notebook which uses the latest trained CustomVision model to identify and classify Street Signs for all NYC neighborhoods.

TBD: Show top & bottom-5 neighborhoods that contain the most/least number of signs from each category

Group 2 - Taxis

@mv1742 mentioned that their team has trained a model to successfully detect NYC taxis using imagery from Open Street Cam.

TBD: Using a random sample of imagery from NYC points of interests or neighborhoods, count the percentage of yellow taxis. Do this repeatedly to determine an estimated average of yellow taxis in NYC traffic

General

Other

Next Meeting

- Tentative: Varun to treat team to PIZZA (or some other cuisine) in NYC!

Attendees: Varun, Manrique, Karan

vr00n commented 5 years ago

@karansaini282 @mv1742 please add to this if I've missed anything here.