Skip to content

krunal3kapadiya/Purple-Monster-TopCoder-Challenge

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

1 Commit
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Purple-Kernel-Eater-Monster-Ideation

This is task to detect the yellow seeds that will turn into black monster seeds. For more information visit this documentations

Prerequisites

Goal

You can see the sample pair of image below. Goal is to achive the best accuracy for the prediction of which seeds will turn to black.

Day One Day Two Predicted Image
image one image two image two

How to run the code

To train images and train model run following code

python train.py or

python train.py --datasetPath 'PATH_OF_DATASET' --outputPath 'PATH_OF_OUTPUT_DIR'

In above code both datasetPath and outputPath are optional.

To test and create model with predicted result file run following command

python test.py or

python test.py --datasetPath 'PATH_OF_DATASET' --outputPath 'PATH_OF_OUTPUT_DIR'

Insights

Training data

  • First indexed_images.csv will be generated, it seperated the both 24 hours before and after images by it's name.
  • Training file will crop the circles in cropped folder, both images folder before and after will be created inside cropped folder.
  • Then the seeds are seperated from above folders. In root directory of project two folders will created named as before and after
  • Store the seeds_converted.csv inside folder saved_csv where the seeds that are converted into purple monster will be saved.
  • Now the size and position will be captured in final_data.csv. If the seeds conveted or not that detail will be stored in maindf.csv
  • It's time to train model. Model will be created based on the images.

Testing data

  • File will be cropped and will find the seeds details.
  • Seeds details will act as an input for saved model
  • If the seed detected as monster it will draw ellipse around it.

About

No description, website, or topics provided.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages