This dataset is conducted with 100 icon labels, which are the most common. A total of 142,416 images from 100 classes are used, with each class containing approximately 1,424 images of size 224×224 and *.jpg format. The dataset was split 80% for training, 10% for validation and 10% for testing.
This dataset contains three folders:
- train (113.932 images)
- validation (14.242 images)
- test (14.242 images)
Each of the above folders contains subfolders for different icons wherein the images for respective icons are present.
The icons are structured into classes with the following labels:
- add
- airplane
- alarm
- arrow_down
- arrow_left
- arrow_right
- arrow_up
- attach
- bag
- barcode
- battery
- bluetooth
- bookmark
- brightness
- calculator
- calendar
- call
- camera
- car
- cart
- chart
- check_mark
- clock
- close
- cloud
- computer
- contrast
- credit_card
- crop
- cursor
- cut
- dashboard
- delete
- dollar
- download
- edit
- external_link
- eye
- fab
- fast_forward
- favorite
- file
- filter
- fingerprint
- fire
- flag
- flashlight
- folder
- gift
- globe
- grid
- headphones
- home
- inbox
- info
- laptop
- light_bulb
- link
- location
- lock
- map
- maximize
- megaphone
- menu
- microphone
- minimize
- mobile
- moon
- music
- mute
- notifications
- overflow_menu
- play
- printer
- profile_avatar
- qr_code
- question
- refresh
- reply
- rewind
- save
- search
- send
- settings
- share
- signal
- sort
- tag
- television
- thumbs_up
- ticket
- trash
- trophy
- unlock
- upload
- user
- video_camera
- volume
- warning