Recurrent Shadow Attention Model (RSAM) for shadow removal in high-resolution urban land-cover mapping
- Award ID(s):
- 1737563
- NSF-PAR ID:
- 10203613
- Date Published:
- Journal Name:
- Remote Sensing of Environment
- Volume:
- 247
- Issue:
- C
- ISSN:
- 0034-4257
- Page Range / eLocation ID:
- 111945
- Format(s):
- Medium: X
- Sponsoring Org:
- National Science Foundation
More Like this
-
Abstract Webs are planar graphs with boundary that describe morphisms in a diagrammatic representation category for $\mathfrak{sl}_k$. They are studied extensively by knot theorists because braiding maps provide a categorical way to express link diagrams in terms of webs, producing quantum invariants like the well-known Jones polynomial. One important question in representation theory is to identify the relationships between different bases; coefficients in the change-of-basis matrix often describe combinatorial, algebraic, or geometric quantities (e.g., Kazhdan–Lusztig polynomials). By ”flattening” the braiding maps, webs can also be viewed as the basis elements of a symmetric group representation. In this paper, we define two new combinatorial structures for webs: band diagrams and their one-dimensional projections, shadows, which measure depths of regions inside the web. As an application, we resolve an open conjecture that the change of basis between the so-called Specht basis and web basis of this symmetric group representation is unitriangular for $\mathfrak{sl}_3$-webs ([ 33] and [ 29].) We do this using band diagrams and shadows to construct a new partial order on webs that is a refinement of the usual partial order. In fact, we prove that for $\mathfrak{sl}_2$-webs, our new partial order coincides with the tableau partial order on webs studied by the authors and others [ 12, 17, 29, 33]. We also prove that though the new partial order for $\mathfrak{sl}_3$-webs is a refinement of the previously studied tableau order, the two partial orders do not agree for $\mathfrak{sl}_3$.more » « less
-
This paper proposes a phase-to-depth deep learning model to repair shadow-induced errors for fringe projection profilometry (FPP). The model comprises two hourglass branches that extract information from texture images and phase maps and fuses the information from the two branches by concatenation and weights. The input of the proposed model contains texture images, masks, and unwrapped phase maps, and the ground truth is the depth map from CAD models. A loss function was chosen to consider image details and structural similarity. The training data contain 1200 samples in the verified virtual FPP system. After training, we conduct experiments on the virtual and real-world scanning data, and the results support the model’s effectiveness. The mean absolute error and the root mean squared error are 1.0279 mm and 1.1898 mm on the validation dataset. In addition, we analyze the influence of ambient light intensity on the model’s performance. Low ambient light limits the model’s performance as the model cannot extract valid information from the completely dark shadow regions in texture images. The contribution of each branch network is also investigated. Features from the texture-dominant branch are leveraged as guidance to remedy shadow-induced errors. Information from the phase-dominant branch network makes accurate predictions for the whole object. Our model provides a good reference for repairing shadow-induced errors in the FPP system.more » « less
- Free Publicly Accessible Full Text
-
Accepted Manuscript1.0
- Journal Article:
- https://doi.org/10.1016/j.rse.2020.111945
