0

I am currently working on text image retrieval where basically, given a test query, I should be able to retrieve all the related images. For example- 'children playing in park', should retrieve all relevant images. I do know the evaluation metric to use for this is Precision@k. I also understand what it is.

However, i don't know how to find it since i don't have associated labels on what images are relevant for what input query. I am using Flickr30K dataset.

Nidutt
  • 185
  • 2
  • 17
  • In such cases you need to annotate the relevance labels to build ur gound-truth – Debasis Mar 01 '22 at 18:04
  • Yeah, i figured. I feel there must be a prebuilt ground label since its a very popular dataset and domain. I see a lot of papers on the same. However I couldn't find any ground truth labels anywhere online. – Nidutt Mar 01 '22 at 21:16

0 Answers0