SnapMEAL : Food Image Recognition Prototype

The aim of  this  project  is  to  develop  an  application  for  smartphone which  counts  calories  using  image  recognition  which  helps  in identifying  food.  A  study  has  been  done  on  a  data  set  of  22,000  images  with  22  different  categories with  1000  different  pictures  of  each  category  using  two-layer  Convolution  Neural  Network  that  helps in  identifying  the  food  and  subsequently  retrieves  calories  from  web.  The  attempt  has  been  made  to keep  a  track  of  the  calorie  intake  in  each  meal  during  the  day.  The  application  allows  you  to  click and  share  pictures  on  your  favorite  social  networking  channel  such  as  Instagram  etc  while  providing a  framework  for  calculating  calories  in  the  background  meanwhile  making  it  simple,  exciting  and interactive. This survey will be private and will only be used for study purposes and will be deleted once done.