Supervised Estimation of Dense Optical Flow.

Date of Submission

December 2020

Date of Award

Winter 12-12-2021

Institute Name (Publisher)

Indian Statistical Institute

Document Type

Master's Dissertation

Degree Name

Master of Technology

Subject Name

Computer Science


Electronics and Communication Sciences Unit (ECSU-Kolkata)


Chanda, Bhabatosh (ECSU-Kolkata; ISI)

Abstract (Summary of the Work)

End-to-end trained Convolutional Neural Network (CNN) have significantly advanced the field of computer vision in recent years, particularly high-level vision problems, because of its strong non-linear fitting ability. In context of optical flow, obtaining dense, ground truth per-pixel for real scenes is difficult and thus rarely available. But CNN in recent years demonstrated that dense optical flow estimation can be cast as a learning problem. However, the state of the art with regard to the quality of the flow has still been defined by traditional methods. In this thesis, firstly, we used a compact but effective CNN model, called U-Net, which contains an encoder part and a decoder part and used benchmark datasets: MPI-Sintel, KITTI and Middlebury; for training and evaluation, in a supervised manner. Secondly, we used some traditional energy-based loss function for dense optical flow estimation. Thirdly, we used backward warping with bilinear interpolation to predict first image and build occlusion mask using ground truth flow. Experimental results show that our proposed method is on par with state-of-the-art supervised CNN methods.


ProQuest Collection ID:

Control Number


Creative Commons License

Creative Commons Attribution 4.0 International License
This work is licensed under a Creative Commons Attribution 4.0 International License.


This document is currently not available here.