Depth Estimation by Finding Best Focused Points Using Line Fitting

Abstract

This paper presents a method for depth estimation using image focus based on the linear regression model. For each pixel, we select two datasets based on the maximum value which is calculated using Laplacian operator. Then linear regression model is used to find lines that approximate these datasets. The best fit lines are found using least squares method. After approximating the two lines, their intersection point is calculated and weights are assigned to calculate the new value for the depth map. The proposed method is compared with four depth estimation algorithms.

Depth coded shape from focus

“…In the early work of Nayar and Nakagawa [16,15], a Gaussian model is fitted to the focus responses. A similar approach based on intersecting a line pair, fitted to the falling and rising slopes of the response profiles is presented in [11]. Asif and Choi trained a neural network on the shape of the focus maxima [2].…”

Depth estimation for autonomous robot navigation: A comparative approach

“…Optical flow has extensively been used thus far for obstacle avoidance and centring a robot in corridor-like environments [4], [5], [6], [7]. Optical flow for depth estimation has been used in [8], [9] and a least squares approach for inferring depth appears in [10], [11]. Furthermore, a large amount of effort has been focussed on using total least squares [12], [13], least squares [14], or constrained total least squares methods [15] for calculating the optical flow field.…”