search

ritwik12 / Celestial-bodies-detection

TensorFlow Image Classifier that can be used to classify whether an image is of a Planet (Earth, Mercury, Mars, etc), Galaxy (Spiral, Elliptical, Irregular), Satellites, Comets, Etc.
https://celestial-bodies-detection.herokuapp.com
GNU General Public License v3.0
84 stars 59 forks source link

Add Training Data #35

Open ritwik12 opened 4 years ago

ritwik12 commented 4 years ago

There are few data available for Uranus, Mars and others. Having a good amount of training data increases accuracy also. Uranus is having only 20 images and for this model it should be at least 25 for better efficiency.

kannandreams commented 4 years ago

I worked to add few uranus images. please can you review the pull request and approve. https://github.com/ritwik12/Celestial-bodies-detection/pull/36

ritwik12 commented 4 years ago

@kannandreams Sure, Reviewed Please check

mgrinstein commented 4 years ago

Hi @ritwik12 ! I've added some training data for Mars, Neptune and Uranus (PR #62 for your approval).

ritwik12 commented 4 years ago

@ferreiramonique Thanks!!

varunpusarla commented 4 years ago

Hey. Do you need more training data ? If yes, I'd like to contribute.

ritwik12 commented 4 years ago

@varunpusarla Yes, we do need more but we need authentic and accurate images if you can help with that.

SambhaviPD commented 4 years ago

@ritwik12 - Is this issue still open? I checked the test data folder, I assume images on similar lines should be okay. Approximately how many images are required?

ritwik12 commented 4 years ago

@SambhaviPD There is never a limit to data, the more the images the better it is. But it would be better to have good and clear images only and not the animated or computerized ones.

SambhaviPD commented 4 years ago

@ritwik12 - Sure, let me first grab a few and submit a PR. Let know about the quality and I shall add more based on your feedback.