Summary

This text document compiles a wide-ranging collection of academic proceedings and conference reports related to computer vision, specifically focusing on facial image recovery, model selection, and domain adaptation for generalization. The material ranges from the 2018 WACV, which featured work on generalized zero-shot learning and identity-preserving face recovery, to the 2017 ECCV conference. Several papers in this set utilize kernel feature maps to enhance action recognition or perform supervised domain adaptation on body skeletons. Notable figures including Piotr Koniusz, Fatih Porikli, and Xin Yu have collaborated extensively across multiple venues, contributing to research areas such as museum artifact identification and wearable camera image recognition. While many papers focus on technical innovations like deeper power normalizations and higher-order pooling techniques, others address practical applications like the Museum Exhibit Identification Challenge, demonstrating significant impacts in domain adaptation beyond traditional supervised learning. The compilation highlights the growing interest in utilizing auxiliary facial attributes alongside core face features to improve models that fail on new data or face different domains, with applications extending from general computer vision tasks to specific challenges in wearable image detection.

Title

Piotr Koniusz

Description

Koniusz, Piotr Koniusz, Data61, CSIRO, Australian National University, ANU, UNSW, Principal Research Scientist, Associate Professor

Keywords

vision, conference, learning, computer, recognition, international, shot, pattern, acceptance, rate, code, image, action, detection, processing, graph, papers

NS Lookup

A 50.6.160.125

Dates

Created 2026-04-15

Updated 2026-04-17

Summarized 2026-04-20