1 / 7

Week 3: Web-Assisted Object Detection

Week 3: Web-Assisted Object Detection. Alejandro Torroella. Papers I’ve been reading. Accurate Image Localization Using Google Maps Street Views City Scale GeoSpatial Trajectory Estimation Object Detection with Discriminatively Trained Part Based Models. GIS Formats and programs.

macon
Download Presentation

Week 3: Web-Assisted Object Detection

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. Week 3:Web-Assisted Object Detection Alejandro Torroella

  2. Papers I’ve been reading Accurate Image Localization Using Google Maps Street Views City Scale GeoSpatial Trajectory Estimation Object Detection with Discriminatively Trained Part Based Models

  3. GIS Formats and programs KML & KMZ: Variation of XML, very straight forward for my purposes GeoServer: Open Source community for GIS information Google Earth: Opens KML and KMZ files

  4. Discriminately Trained Deformable Part Based Models Downloaded the code, had to install Linux to run it, ran into some issues with the compiler configuration, eventually got the code to train on PASCAL database, took about half a day to do so. Need to create my own database of outside objects (hydrants, lamp posts, bus stops, ATMs, etc.) and train models on them.

  5. Perhaps look into F.R.E.A.k. for object detection? DPMs don’t take into account orientation, models take a long time to train, results might still be lousy even with assistance of GIS data. F.R.E.A.K. is more robust, super fast, doesn’t require extensive training, and results could be more promising even without assistance of GIS data.

  6. Goals for next week Look into data fusion methods Obtain a database of images of different outside objects Train DPM on these databases Perhaps do more research on better object detection methods

  7. Thank youFin.

More Related