This dataset comprises approximately 5,000 visual records of pre-Hispanic Peruvian ceramic vessels from the Paracas, Moche, and Nazca cultures. The imagery was compiled via automated web scraping from the public digital repositories of the Ministry of Culture of Peru (Ministerio de Cultura del Perú), conducted with their express institutional authorization and institutional consent for academic research purposes. It is specifically designed to support the training, evaluation, and benchmarking of deep learning and computer vision models focused on digital artifact restoration, inpainting, and surface reconstruction. Dataset Characteristics Data Provenance & Authorization: Collected directly from official, public archaeological repositories curated by the Ministry of Culture of Peru under formal permission. Visual records reflect standardized museum documentation across multiple regional collections. Cultural Coverage: Features archaeological pottery spanning three prominent pre-Columbian traditions: Paracas (known for distinct polychrome post-fire painting and incised designs), Moche (characterized by detailed sculptural shapes and bichrome fine-line iconography), and Nazca (distinguished by vibrant polychrome pre-fire slip painting). Condition States: Contains images documenting ceramics in diverse states of preservation, ranging from intact, well-preserved pieces to degraded artifacts exhibiting fractures, surface abrasion, pigment loss, cracking, and missing structural sections. Unpaired Structure: Visual samples are structured predominantly as unpaired (unaligned) domain sets rather than matched before-and-after pairs. This layout makes the dataset directly suitable for unsupervised image-to-image translation architectures (such as CycleGAN, DualGAN, or diffusion-based unpaired frameworks) and self-supervised inpainting pipelines.
David Senyi Sen Prado, Edwin Jonathan Escobedo Cárdenas· Zenodo (CERN European Organi...· 0 citations
This dataset comprises approximately 5,000 visual records of pre-Hispanic Peruvian ceramic vessels from the Paracas, Moche, and Nazca cultures. The imagery was compiled via automated web scraping from the public digital repositories of the Ministry of Culture of Peru (Ministerio de Cultura del Perú), conducted with their express institutional authorization and institutional consent for academic research purposes. It is specifically designed to support the training, evaluation, and benchmarking of deep learning and computer vision models focused on digital artifact restoration, inpainting, and surface reconstruction. Dataset Characteristics Data Provenance & Authorization: Collected directly from official, public archaeological repositories curated by the Ministry of Culture of Peru under formal permission. Visual records reflect standardized museum documentation across multiple regional collections. Cultural Coverage: Features archaeological pottery spanning three prominent pre-Columbian traditions: Paracas (known for distinct polychrome post-fire painting and incised designs), Moche (characterized by detailed sculptural shapes and bichrome fine-line iconography), and Nazca (distinguished by vibrant polychrome pre-fire slip painting). Condition States: Contains images documenting ceramics in diverse states of preservation, ranging from intact, well-preserved pieces to degraded artifacts exhibiting fractures, surface abrasion, pigment loss, cracking, and missing structural sections. Unpaired Structure: Visual samples are structured predominantly as unpaired (unaligned) domain sets rather than matched before-and-after pairs. This layout makes the dataset directly suitable for unsupervised image-to-image translation architectures (such as CycleGAN, DualGAN, or diffusion-based unpaired frameworks) and self-supervised inpainting pipelines.
David Senyi Sen Prado, Edwin Jonathan Escobedo Cárdenas· Zenodo (CERN European Organi...· 0 citations
We use cookies to run the site and, with your consent, for analytics and to show ads.
See our Cookie Policy.
