2020/07/14 by Mark Hamilton, Hamilton, Mark, Stephanie Fu +22 · 4 citations
Computer Science · Mathematics · #Artificial intelligence #Computer Vision and Pattern Recognition (cs.CV) #Computer science #Data mining #Domain (mathematical analysis) #FOS: Computer and information sciences #Generative Adversarial Networks and Image Synthesis #Generative grammar #Geography #Graphics (cs.GR) #Image (mathematics) #Image Retrieval and Classification Techniques #Image retrieval #Information Retrieval (cs.IR) #Information retrieval #Machine Learning (cs.LG) #Machine Learning (stat.ML) #Mathematics #Mosaic #Music and Audio Processing #Parametric statistics #Pattern recognition (psychology) #Similarity (geometry) #Span (engineering) #Video Analysis and Summarization #cs.CV #cs.GR #cs.IR #cs.LG #stat.ML
paper · pdf · doi:10.48550/arxiv.2007.07177
published in arXiv (Cornell University), 133-155 (Cornell University)
openalex publication_date 2020/07/14 · arxiv created 2021/02/28 · arxiv updated 2021/03/02 · openalex created_date 2022/07/26 · openalex updated_date 2026/08/08
We introduce MosAIc, an interactive web app that allows users to find pairs\nof semantically related artworks that span different cultures, media, and\nmillennia. To create this application, we introduce Conditional Image Retrieval\n(CIR) which combines visual similarity search with user supplied filters or\n"conditions". This technique allows one to find pairs of similar images that\nspan distinct subsets of the image corpus. We provide a generic way to adapt\nexisting image retrieval data-structures to this new domain and provide\ntheoretical bounds on our approach's efficiency. To quantify the performance of\nCIR systems, we introduce new datasets for evaluating CIR methods and show that\nCIR performs non-parametric style transfer. Finally, we demonstrate that our\nCIR data-structures can identify "blind spots" in Generative Adversarial\nNetworks (GAN) where they fail to properly model the true data distribution.\n