Skip to content

UserGuide.md

Latest commit

 

History

History
46 lines (34 loc) · 3.24 KB

UserGuide.md

File metadata and controls

46 lines (34 loc) · 3.24 KB

VISE User Guide

This guide assumes that you have already installed VISE and it is running.

  1. Open http://localhost:9971 in a web browser as shown in this screenshot
  • Do not open localhost:9971 in multiple browser windows. Just use a single browser window
  • Once you have trained a search engine or loaded a search engine, you cannot load or train another search engine unless you restart (stop + start) VISE through Kitematic and revisit http://localhost:9971.

  1. Create a new search engine by typing in the search engine name and pressing the Create button as shown in this screenshot

  2. You will now be taken to the Setting page. Copy all your images to the local folder that maps to the docker folder /opt/ox/vgg/mydata/images. For MacOS and Windows users, this folder is the local folder that was mapped to /opt/ox/vgg/mydata/images as described in Step 4 of the installation guide for MacOS and Windows.

  3. Now you will see an information page which contains the disk-space, memory and time required to make the images searchable as shown in this screenshot. If you are happy with these details, proceed ahead by pressing Proceed.

  4. The main process of making images searchable begins now. This process involves going through different stages: Preprocessing, Stage-1 ... Stage-5. as shown in these screenshots:

  1. Once this process completed, you will be automatically forwarded to this URL http://localhost:9971 as shown in this screenshot -- where the list of images contain your images.
  • Click on any of the images shown in the list
  • Using mouse, define a rectangular region in the selected image (click and drag to draw a region)
  • Press Search button at the bottom of the page
  • Now, you will see a list of images ranked according to a score (higher score for better matching results)
  • Click Detailed matches to see the details of matched image
  • You can compare the query image region with region in search result using Image Comparison link at the bottom of this page

Known Issues

Last updated: 06 Sep. 2017
Abhishek Dutta