Abstract
Feature specific imaging is a computational imaging technique that minimizes the number of measurements needed to sufficiently reconstruct a scene by using a priori knowledge (e.g., the scene’s second-order statistics) to judiciously, as well as possibly adaptively, choose the projection vectors to be measured. Here, we have developed an approach to three-dimensional adaptive feature specific imaging that takes into account the obstruction of distant objects by closer objects in the adaption of the projections and in the reconstruction algorithm. The developed system reconstructs the cross-range image of the scene at each range bin from a set of range resolved measurements from all the return from the scene at that range using only a single photodetector, while adapting to the obstruction of the scene by closer objects. Simulations and a proof-of-concept demonstration of adaptive three-dimensional feature specific imaging are presented.
Highlights
Computational imaging is a means of reducing the number of measurements needed to sense and reconstruct an image of a sparse scene, whether it be an image in space,[1,2] time,[3] or delay.[4]
One goal of 3-D adaptive FSI6 (AFSI) algorithm is to develop an adapted training set at each range AðkiÞ 1⁄4 faðki;Þjg that represents the obscured training scenes that would be seen by the projector of the basis vectors and the single-pixel measurement system due to obstructions by xk[0] for all k 0 < k, where i represents the iteration index number
If the goal was to estimate the unseen portions of the objects, using the algorithm without adaptation for obstruction (AFO) might be useful, but it is likely that the results of the algorithm with AFO could be used to estimate the continuation of the obstructed objects with high fidelity
Summary
Computational imaging is a means of reducing the number of measurements needed to sense and reconstruct an image of a sparse scene, whether it be an image in space,[1,2] time,[3] or delay.[4]. A training set of faces with a wide range of different eyes, noses, mouths, and chins, with or without hats and glasses can be used to faithfully reconstruct the face of a person not used in the training set.[6] Though the image of the actual face to be reconstructed is not in the training set of faces, it is assumed that the various features in the actual Another method is to use a predetermined basis that is known to be compressive, e.g., the Hadamard basis, discrete cosine transform (DCT) basis, or a wavelet basis. The obstruction of the background objects (faces) will result in a significant altering of the preferred ordering of the basis vectors needed to efficiently reconstruct the background faces, along with the unobstructed foreground objects (faces). These adapted training sets are used in subsequent choices of projection vectors with the goal of effectively reconstructing the foreground and background objects with a limited number of projections
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
