Silhouette Coherence for Camera Calibration under Circular Motion

Abstract

Abstract-We present a new approach to camera calibration as a part of a complete and practical system to recover digital copies of sculpture from uncalibrated image sequences taken under turntable motion. In this paper, we introduce the concept of the silhouette coherence of a set of silhouettes generated by a 3D object. We show how the maximization of the silhouette coherence can be exploited to recover the camera poses and focal length. Silhouette coherence can be considered as a generalization of the well-known epipolar tangency constraint for calculating motion from silhouettes or outlines alone. Further, silhouette coherence exploits all the geometric information encoded in the silhouette (not just at epipolar tangency points) and can be used in many practical situations where point correspondences or outer epipolar tangents are unavailable. We present an algorithm for exploiting silhouette coherence to efficiently and reliably estimate camera motion. We use this algorithm to reconstruct very high quality 3D models from uncalibrated circular motion sequences, even when epipolar tangency points are not available or the silhouettes are truncated. The algorithm has been integrated into a practical system and has been tested on more than 50 uncalibrated sequences to produce high quality photo-realistic models. Three illustrative examples are included in this paper. The algorithm is also evaluated quantitatively by comparing it to a state-of-the-art system that exploits only epipolar tangents.

Shape from Photographs: A Multi-view Stereo Pipeline

“…The image resolution is 5 Mpix and the camera motion was recovered by standard structure from motion techniques [55] and further refined using a silhouette-based technique [21]. The object exhibits significant self-occlusions, a large concavity in the chest and two thin legs which make it a very challenging test to validate the new ballooning term.…”

Hardware-based camera calibration and 3D modelling under circular motion

“…Hernandez et al [18] extended the epipolar tangency criteria to silhouette coherence and provided a more generic method to compute contour coherence under circular motion. Although they used image contours instead of silhouettes to accelerate the optimization speed, the coherence measure was complicated and changed, which could affect its robustness and accuracy.…”

“…During the image capture process, we rotate the turntable while keeping the camera fixed. Similar to the work of [18], under such circular motion constraints, we have two different kinds of camera parameters to calibrate, which we define as hard camera parameters and soft camera parameters.…”

Automatic 3D Object Segmentation in Multiple Views using Volumetric Graph-Cuts

“…In order to reduce the demands placed on the user we aim to exploit the constraints that exist within the image sequence. Note that the sequence contains a series of multiple views of the same rigid 3D object, therefore the segmentations must satisfy a silhouette coherency constraint [7]. This constraint follows from the knowledge that the images are formed from projections of the same rigid 3D object with the correct segmentations being the corresponding silhouettes.…”

“…Silhouette coherency has previously been used for camera calibration [7] under the further constraint of circular motion. Here we use the inverse approach and assume that given accurate calibration we can propagate knowledge across multiple views as segmentation proceeds.…”