0

I have built an image search using VGG16 engine, I have a data set of about 20,000 images. It takes 4 mins to go through the search. what are the techniques that I can use to shorten this time?

Duck Dodgers
  • 3,409
  • 8
  • 29
  • 43
Himanshu Bansal
  • 1
  • Use approximate nearest neighbour search library such as https://github.com/spotify/annoy – Zaw Lin Jan 18 '19 at 22:01
  • Providing additional details like how are you running your test? What framework are you using (tensorflow, pytorch)? And your system configuration might allow others to understand your situation little better. Check this link for general guidelines on posting questions. https://stackoverflow.com/help/how-to-ask . – Mohana Rao Jan 28 '19 at 21:20
  • i am using keras with tensorflow backend. – Himanshu Bansal Jan 30 '19 at 05:23

1 Answers1

0

I have reduced the time latency to about one fourth of the previous one. I can be done by using a classifier on the images prior to finding the eucledian distance between the embeddings of the images.I used SVM and trained it with about 10 different categories , the average latency was thus reduced to about one forth because of decrease in the number of comparisons required.

Himanshu Bansal
  • 1