# How to recognize multiple objects in an image

I'm trying to do multi-object detection and recognition using Mathematica.

For example, in this image:

img = Import["https://i.stack.imgur.com/lbSwV.jpg"];


the result of recognition should be a baby and a cat.

But when I use ImageIdentify, the result is

Entity["Concept", "DomesticCat::jpx55"]


Any ideas for how to recognize both the baby and the cat?

P.S. what I'm working on is a more general image set which contains various types of objects, such as men, women, babies, cats, dogs, cars, apples, bananas, etc. What I need is a tool or method which can recognize all such objects in an image.

• I think with the current version 11 there is no single one-command to recognize multiple objects. Currently, it recognizes cats more than boys because the model is trained with more cat pictures than boys... All identification tasks are single class at a time. May be with the new release it might be possible when there is RNN. – s.s.o Jan 23 '17 at 21:00
• @s.s.o Thank you so much. But I think I can still use a series of different commands of image processing and machine learning to build a model for multi-object recognition. Am I right? – Chang Liu Jan 24 '17 at 8:17
• – Payas Pandey Jan 24 '18 at 10:41
• @PayasPandey your input is valuable but it does not really answer the question about Mathematica so I moved it to comments. – Kuba Jan 24 '18 at 11:00