C
Carlos D. Castillo
Researcher at University of Maryland, College Park
Publications - 106
Citations - 5510
Carlos D. Castillo is an academic researcher from University of Maryland, College Park. The author has contributed to research in topics: Facial recognition system & Convolutional neural network. The author has an hindex of 31, co-authored 98 publications receiving 4111 citations. Previous affiliations of Carlos D. Castillo include University of Chile & Simón Bolívar University.
Papers
More filters
Proceedings ArticleDOI
Activity Detection in Untrimmed Videos Using Chunk-based Classifiers
Joshua Gleason,Steven Schwarcz,Rajeev Ranjan,Carlos D. Castillo,Jun-Cheng Chen,Rama Chellappa +5 more
TL;DR: An algorithm which is based on the proposition that despite the differences between activity classification and detection, a strong classifier can still be used to achieve state-of-the-art performance in detection by breaking the video into multiple overlapping chunks and classifying each individually is proposed.
Posted Content
Biconvex Relaxation for Semidefinite Programming in Computer Vision
TL;DR: The biconvex relaxation (BCR) algorithm as mentioned in this paper transforms a general SDP into a specific bicon vex optimization problem, which can then be solved in the original, low-dimensional variable space at low complexity.
Proceedings ArticleDOI
Video-Based Face Association and Identification
TL;DR: Experimental results on the newly released JANUS challenge set 3 (JANUS CS3) dataset show that the proposed target face association (TFA) technique generates robust representations from target-annotated videos and demonstrates good performance for the task of video-based face identification problem.
Proceedings ArticleDOI
Uncertainty Modeling of Contextual-Connections Between Tracklets for Unconstrained Video-Based Face Recognition
TL;DR: In this article, the uncertainty-gated graph (UGG) is proposed to propagate the identity from high-quality faces to low-quality ones through contextual connections, which are constructed based on context such as body appearance.
Proceedings ArticleDOI
Where in the World is this Image? Transformer-based Geo-localization in the Wild
TL;DR: This work proposes TransLocator, a unified dual-branch transformer network that attends to tiny details over the entire image and produces robust feature representation under extreme appearance variations and is found to be more effective than previous methods.