For a small project for multi-modal semantics, I am looking for a dataset which contains textual and visual information about different objects i.e. cars, flowers etc. Requirements Categories - More than 20 objects Type - jpg, png

  • 1
    visual information Does that mean you need images? If so, in what format. Any size requirements? And what is the textual information you want? How much data? It also helps if you give some background what you need them for. Please read this, then edit and make your question specific, it may get closed in its current form. – user4293 Jan 25 '17 at 10:24
  • I have made the required changes @JanDoggen – cpwah Jan 25 '17 at 10:37

Fortunately, there are a lot of free datasets published for the field of "computer vision". Based on the description of your needs, you should maybe take a look at ImageNet.

  • I have already seen ImageNET however it does not provide a description of the objects. – cpwah Feb 1 '17 at 9:25
  • If the description you want can be a list of labels, you can try Open Images. Here is a viewer to browse the categories : http://openimages.oldjpg.com/ – Ettore Rizza Feb 1 '17 at 10:10
  • Label can be use to represent a concept or a category which is different from the description. Below I mentioned couple of interesting dataset which I find for my work. We are also going to publish our dataset with our publication – cpwah Feb 1 '17 at 10:21

Based on the requirements, following datasets can be use.

Pinterest Multimodal Dataset ToolBox

Text and Vision (TVGraz) Dataset

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.