Skip to content

Latest commit

 

History

History
49 lines (42 loc) · 1.51 KB

README.md

File metadata and controls

49 lines (42 loc) · 1.51 KB
Raw

Image-Scene-Classification

Competition Link: https://competitions.codalab.org/competitions/28113

Data overview

We are employing a newly collected Camera Scene Detection dataset consisting of images belonging to 30 different classes:

  • Portrait
  • Group Portrait
  • Kids / Infants
  • Dogs
  • Cats
  • Macro / Close-up
  • Food / Gourmet
  • Beach
  • Mountains
  • Waterfall
  • Snow
  • Landscape
  • Underwater
  • Architecture
  • Sunrise / Sunset
  • Blue Sky
  • Overcast / Cloudy Sky
  • Greenery / Grass
  • Autumn Plants
  • Flower
  • Night
  • Shot Stage / Concert
  • Fireworks
  • Candle light
  • Neon Lights / Neon Signs
  • Indoor
  • Backlight / Contre-jour
  • Text / Document
  • QR Code
  • Monitor Screen

The dataset is divided into:

train data: 9897 images of resolution 576 x 384 px from the above 30 classes that can be used for training the model. validation data: 600 images of resolution 576 x 384 px provided from the beginning of the challenge and are meant for the participants to get online feedback from the validation server.
test data: 600 images of resolution 576 x 384 px; the participants will get access to them when the final evaluation phase starts and the results will be announced after the challenge is over and the winners are decided.

Data access

Development phase / Learning:

Train Data (images divided by classes) - http://data.vision.ee.ethz.ch/ihnatova/camera_scene_detection_train.zip
Validation Data (only input images) - https://data.vision.ee.ethz.ch/ihnatova/camera_scene_detection_validation.zip