Auto-Cropping: Image Background Removal using Deep Learning
Background removal is a task that is quite easy to do manually, or semi manually (Photoshop, and even Power Point has such tools) if you use some kind of a “marker” and edge detection. However, fully automated background removal is quite a challenging task.So i was quite inspired by this idea.So I thought to work with this in a Secure and private AI project showcase challenge.
I used several techniques to perform segmentation but with Mask-RCNN pretrained model I was not able to get a proper segementation.
Finally i studied about DeepLab is a state-of-art deep learning model for semantic image segmentaion it help me achieving my desired results for Demo.I used pretrined model for this project 'xception_coco_voctrainaug' on COCO dataset.
How To run
Script: Crop.py
arguments : image_path ,background_path
Example:python Crop.py --image_path="/home/usman/Desktop/1936314_1690334921205554_3960288358644789045_n.jpg" -- background_path="/home/usman/Desktop/pexels-photo-949587.jpeg"
Explanation :
--image_path : path to the image which need to be cropped
--background_path :path to the background that you need to paste your cropped image on
Visual Example:
Path to input image and the background image
In the following directories you can see your cropped image and the pasted image(on a background)
cropped_image directory
pasted_image directory
Helping material
Material that helped me and will help me in the future
https://arxiv.org/abs/1606.00915
https://arxiv.org/abs/1505.04597
https://arxiv.org/abs/1703.03872
https://arxiv.org/abs/1703.06870
https://medium.com/nanonets/how-to-do-image-segmentation-using-deep-learning-c673cc5862ef
https://medium.com/fnplus/blue-or-green-screen-effect-with-open-cv-chroma-keying-94d4a6ab2743
https://towardsdatascience.com/background-removal-with-deep-learning-c4f2104b3157