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 57 forks source link

Add Pluto and other Dwarf planets #88

Open ritwik12 opened 2 years ago

ritwik12 commented 2 years ago

Currently we only have 8 planets and it would be great to include Pluto and other dwarf planets too.

glunkad commented 2 years ago

Hey @ritwik12 , I'm interested in contributing to this issue, so before I start working on it, would you mind sparing your time explaining what the issue is about and pointing me to some resources to get started.

ritwik12 commented 2 years ago

Yeah. so just like we have 8 planets right now for the classification. Our model is not trained for Pluto and Dwarf planets. If we try to give an image of Pluto to our model. It will classify it as wrong.

ritwik12 commented 2 years ago

Need to add Training images for Pluto, then add test images. Train the model based on those images.

SSahas commented 1 year ago

Hello @ritwik12 ,

I'm interested in solving this issue, and I cloned this repository and I collected some pluto images and added them to both training and test data and added "pluto" to retrained_labels.txt and label_image.py and now how to start training ??is this the correct way ??

ritwik12 commented 1 year ago

@SSahas Please follow the readme.md. I think you also don't need to update retrained_labels.txt as that is something done on its own with this command below.

To train model you need to run python retrain.py --bottleneck_dir=bottlenecks --how_many_training_steps=500 --model_dir=inception --summaries_dir=training_summaries/basic --output_graph=retrained_graph.pb --output_labels=retrained_labels.txt --image_dir=./training_data

SSahas commented 1 year ago

hey @ritwik12 , I trained the model for classifying Pluto planet and got the final test accuracy : 87% , I also run the run.py and the webapp is classifying Pluto images correctly but the retrained_graph.pb file is greater than 50 mb so I am not able to push it , can I use git LFS now??

ritwik12 commented 1 year ago

@SSahas You meant retrained_graph.pb? The current one in repo is already 83MB

SSahas commented 1 year ago

hey @ritwik12 , I opened a pull request called "Train the model for detecting pluto". I trained the model and the flask app can also able to detect the images as pluto . please look in to it.

please tell if there are any things to be changed.

the-silent-geek commented 9 months ago

Hey @ritwik12 , I'm interested in contributing to this issue. This is my first time contributing. I want to add a dwarf planet named 'Makemake' to the classification.