How to categorize paper / plastic image is 2D or 3D object in image ?

I am developing a program to categorise image of paper or plastic object, (Flat or scrambled paper). After taking image , i need to collect the data whether the object in image is 2D OR 3D.

[Image Analyst]

You forgot to attach the image so we have no idea what to answer except to suggest you use the proper sensor, like one that can detect heights/distances/topography, like a profilometer, kinect, rotating white light scanner, or something similar.

Please read this

[Hassan Mehmood Khan]

Sir, i do not use any sensor. Can i use webcam only and take images then use their properties to detect 2d or 3d. I have attached some samples.

[Image Analyst]

Try asking regionprops() for the solidity. By the way, a webcam is a sensor, though not a 3-D one.

[OCDER]

Are you planning to use deep learning? There are a lot of clever artists who can draw 3D-looking objects on a flat paper. So is that 2D or 3D? If you take a picture of a picture of a 3D crumbled paper (for instance, a screenshot of a crumbled paper) is that 3D or 2D? You could use the camera sensor to take images at two different angles and see if the object shape changes accordingly to a crumbled paper or a flat paper. Search "Stereo Imaging" on the web.

[Hassan Mehmood Khan]

Sir i am not using deep learning. Just want to Use SVM for classification. Need to classify crumbled paper, Crumbled plastic, Flat Paper and Tin Can. Need help to identify what type of features i can use for classification.

Answer 1

[Image Analyst]

Try stdfilt() and look for uniform areas (flat paper) or rough areas (crumpled paper). Attach your images if you need any more help.