More Like this

1. Adaptive display images

   https://doi.org/10.1142/s0219530519410112  

   Zhi, Meipeng; Xu, Yuesheng (January 2020, Analysis and Applications)

   We develop a numerical method for construction of an adaptive display image from a given display image which is an artificial scene displayed in a computer screen. The adaptive display image is encoded on an adaptive pixel mesh obtained by a merging scheme from the original pixel mesh. The cardinality of the adaptive pixel mesh is significantly less than that of the original pixel mesh. The resulting adaptive display image is the best [Formula: see text] piecewise constant approximation of the original display image. Under the assumption that a natural image, the real scene that we see, belongs to a Besov space, we provide the optimal [Formula: see text] error estimate between the adaptive display image and its original natural image. Experimental results are presented to demonstrate the visual quality, the approximation accuracy and the computational complexity of the adaptive display image. 

   more » « less
2. Constitutive Relations From Images

   https://doi.org/10.1115/1.4068870  

   Wihardja, Adeline; Bhattacharya, Kaushik (August 2025, Journal of Applied Mechanics)

   Abstract Constitutive relations close the balance laws of continuum mechanics and serve as a surrogate for a material in the design and engineering process. The problem of obtaining the constitutive relations is an indirect inverse problem where both the relation and the quantities that define the relation have to be inferred from experimental observations. The advent of full-field observation techniques promises a new ability to learn constitutive relations under realistic operational conditions. However, this is done in two steps, first by obtaining deformations from the images and then by obtaining the constitutive relation from deformations and forces. This leads to a variety of difficulties. In this article, we propose a novel approach that enables us to obtain constitutive relations directly from the raw data consisting of images and force measurements. 

   more » « less
3. Extractors for Images of Varieties

   https://doi.org/10.1145/3564246.3585109  

   Guo, Zeyu; Volk, Ben Lee; Jalan, Akhil; Zuckerman, David (June 2023, Conference proceedings of the annual ACM Symposium on Theory of Computing)

   We construct explicit deterministic extractors for polynomial images of varieties, that is, distributions sampled by applying a low-degree polynomial map 𝑓 to an element sampled uniformly at random from a 𝑘-dimensional variety 𝑉. This class of sources generalizes both polynomial sources, studied by Dvir, Gabizon and Wigderson (FOCS 2007, Comput. Complex. 2009), and variety sources, studied by Dvir (CCC 2009, Comput. Complex. 2012). Assuming certain natural non-degeneracy conditions on the map 𝑓 and the variety 𝑉 , which in particular ensure that the source has enough min-entropy, we extract almost all the min-entropy of the distribution. Unlike the Dvir–Gabizon–Wigderson and Dvir results, our construction works over large enough finite fields of arbitrary characteristic. One key part of our construction is an improved deterministic rank extractor for varieties. As a by-product, we obtain explicit Noether normalization lemmas for affine varieties and affine algebras. Additionally, we generalize a construction of affine extractors with exponentially small error due to Bourgain, Dvir and Leeman (Comput. Complex. 2016) by extending it to all finite prime fields of quasipolynomial size. 

   more » « less
4. Attributes in Multiple Facial Images

   https://doi.org/10.1109/FG.2018.00053  

   Liu, Xudong; Guo, Guodong (May 2018, IEEE International Conference on Automatic Face & Gesture Recognition (FG 2018))

   Facial attribute recognition is conventionally computed from a single image. In practice, each subject may have multiple face images. Taking the eye size as an example, it should not change, but it may have different estimation in multiple images, which would make a negative impact on face recognition. Thus, how to compute these attributes corresponding to each subject rather than each single image is a profound work. To address this question, we deploy deep training for facial attributes prediction, and we explore the inconsistency issue among the attributes computed from each single image. Then, we develop two approaches to address the inconsistency issue. Experimental results show that the proposed methods can handle facial attribute estimation on either multiple still images or video frames, and can correct the incorrectly annotated labels. The experiments are conducted on two large public databases with annotations of facial attributes. 

   more » « less
5. SpoofGAN: Synthetic Fingerprint Spoof Images

   https://doi.org/10.1109/TIFS.2022.3227762  

   Grosz, Steven A.; Jain, Anil K. (January 2023, IEEE Transactions on Information Forensics and Security)