Please enable JavaScript.
Coggle requires JavaScript to display documents.
Mobile Multi-Food Recognition Using Deep Learning (protocol (Bounding…
Mobile Multi-Food Recognition Using Deep Learning
Upgrade from previous
multi-object recognition
protocol
Bounding circle
remove background
user drawn
whole plate
not signle food items
send to cloud
amazon
Elastic map reduce
task distribution
master node
slaves
single output
Region collection
algorithm
selective search
grouping pixels
sub-units of food
similarities
color
RGB
HSV
LAB
texture
hog
size
Region description
Caffe deep-learning framework
feature vector
Region mining
positive region
includes target
Deep learning
Food name prediction
prediction tags
Experimental results
FODD database
30 categories
multiple food instances
93% recognition
Volume estimation
users finger