Copyright © 2014 by Intelligent Systems Laboratory, Computer Science Department, Technion - Israel Institute of Technology, Haifa 3200003, Israel. All rights reserved


TheGrue.org

Using Deep Learning to Estimate the Gradient and Curvature of Image Patches

Course: 236754 (Project in intelligent systems) or 236874 (Project in computer vision) (3 points)
Instructors: Prof. Micha Lindenbaum and Elad Osherov
Number of students: 2
prerequisites: Computer Vision and/or Image Processing 

 

Gradient and curvature information is a basic information in an image, required in many computer vision tasks, such as object classification, edge detection, segmentation, etc.

In this project you will design a deep neural network, which operates over image patches to estimate the gradient and curvature information a patch holds. This network would provide a continuous regression model for several scales of image patches. Such information can optionally be used in tasks such as image completion or image generation.
The project can be implemented in several Deep learning environments such as MatConvNet, TensorFlow or Torch. We will provide the students an access to a computation server equipped with GPUs.
The students would preferably have at least basic knowledge in machine learning and computer vision/image processing.

Using Deep Learning to Estimate the Gradient and Curvature of Image Patches

 

Contact information:

Elad Osherov
Email: This email address is being protected from spambots. You need JavaScript enabled to view it.