# Optical Flow Problems

I was wondering if any of you guys could help me with the next 3 problems.

They're about estimating the lamda on the Optical Flow equation $E = E_s + λE_c$, for different situations that try to estimate pixel motion in videos.

1) Suppose we're trying to estimate motions in a video gotten from a drone. Do we have to use a small, medium, or large value for $λ$? You need to justify what you answer by using the integral for $E$, $E_s$, and $E_c$.

2) Now we try to estimate some smooth motions. Do we have to use a small, medium, or large value for $λ$? You need to justify what you answer by using the integral for $E$, $E_s$, and $E_c$.

3) After applying the optical-flow method, we see that estimated motions are parallel to the (local) image gradient. To fix this, do we have to use a small, medium, or large value for $λ$? You need to justify what you answer by using the integral for $E$, $E_s$, and $E_c$.

Thanks!

• Some resent finding in the area of Optical Flow: In the new data-set for Optical flow you can find several recent methods in this area. The website URL is "sintel.is.tue.mpg.de" And the some of papers you should read. FlowFields C. Bailer, B. Taetz and D. Stricker. Flow Fields: Dense Correspondence Fields for Highly Accurate Large Displacement Optical Flow Estimation. ICCV 2015. DiscreteFlow M. Menze, C. Heipke and A. Geiger. Discrete Optimization for Optical Flow. GCPR 2015 EpicFlow J. Revaud, P. Weinzaepfel, Z. Harchaoui and C. Schmid. EpicFlow: Edge-Preserving Interpolation of Correspondenc – Farshid PirahanSiah Oct 8 '15 at 8:24
• This looks very much like homework. Can you please write a little bit more about the problem and especially what have you tried and what exact difficulties are you facing? – A_A Mar 6 '16 at 13:46