[Paper] Efficient Acquisition of Light Transport Matrix Using Push Broom-Type Projector Illumination

[Paper] Radiometric Compensation Using Color-Mixing Matrix Reformed from Light Transport Matrix

“…We follow that aim in this study, and apply projector illumination used in existing methods to the design of the sensing matrix. In our previous work 27) , we adopted push broom-type projector illumination, i.e. verticalstrip illuminations with Δw × N pixels and horizontalstrip illuminations with N × Δw pixels, where Δw indicates the width of each narrow illumination that satisfies ΔwL = N .…”

“…Since BFS can simply and accurately measure the geometric correspondences, it is compared with the acquisition method of light transport 7) . According to our previous work 27) , push broom-type scanning (PBS) approximately find direct/indirect light transports from 3D objects, by computing the logical product of vertical/horizontal-strip camera responses, that is, PBS handles the same 2L-kinds of camera responses as CS to acquire the light transport matrix. We compare the 3M 2 ×3(N/4) 2 light transport matrices yielded by BFS with 4 × 4 pixels and PBS with Δw = 4 pixels to that yielded by CS.…”

Dense Light Transport for Relighting Computation Using Orthogonal Illumination Based on Walsh-Hadamard Matrix

“…In this subsection, we examine the impact of this color crosstalk. We also demonstrate BFS with 4 × 4 pixels and push broom-type scanning (PBS) with a width of 4 pixels [26], for comparisons with our proposal. The system outputs individually color illumination from β-channel, β ∈ {r, g, b}, in the projector, and it sequentially captures the corresponding camera responses for each color illumination according to each method.…”

“…where we use K-sets of projector illumination and the corresponding camera response [26]. Since the projector illumination vectors have (P (l) β ) T P (k) β = δ lk , where δ lk indicates the Kronecker delta, we can extract the k-th camera response vector C (k) αβ , which corresponds to projector vector P (k) β , from matrix T αβ by computing,…”

“…Based on this idea, we propose a practical method that simultaneously acquires the direct and global light transports to obtain a dense and high-rank light transport matrix; it supports a wide variety of relighting through the use of photographic, artistic, and complex illuminations. Acquisition of a light transport matrix involves designing a set of projector illuminations and storing the corresponding camera responses in a sequence [26]. We employ the projector illumination patterns (orthogonal illumination) generated by a Walsh-Hadamard matrix, in which one half of the elements are used for positive pattern and the other half are used for negative pattern.…”