Tieniu Tan

TL;DR: In this paper, an image stitching testing method based on illumination direction inconsistency was proposed, which comprises the following steps: selecting a pair of target human faces from a to-be-tested image, and building three-dimensional human face models fitting the target human face images; aligning the 3D face models with 2D human face image images on the to be tested image; using images generated through rendering under light of different light intensity to obtain reflection transfer coefficients at sample points of the three dimensional human face face models through fitting; and comparing the two illumination coefficients, and obtaining a

TL;DR: Zhang et al. as mentioned in this paper proposed a graph attentive relational network (GARN) to learn aligned image-text representations by modeling the relationships between noun phrases in a text for the identity-aware imagetext matching.

TL;DR: This paper proposes exploring multiple order local binary patterns (MOLBP) as features for learning, and develops a localized multi-boost learning (LMBL) algorithm to combine the different features for classification.

TL;DR: This paper introduces relation class labels into conventional high-order multiplicative interactions with pairwise input samples, and proposes a conditional high- order Boltzmann Machine (CHBM), which can learn to classify the data relation in a binary classification way.

TL;DR: Wang et al. as discussed by the authors proposed a scale-adaptive heatmap regression (SAHR) method, which can adaptively adjust the standard deviation for each keypoint, which is more tolerant of various human scales and labeling ambiguities.