Autoencoder, as an essential part of many anomaly detection methods, is lacking flexibility on normal data in complex datasets. U-Net is proved to be effective for this purpose but overfits on the training data if trained by just using reconstruction error similar to other AE-based frameworks. Puzzle-solving, as a pretext task of self-supervised learning (SSL) methods, has earlier proved its ability in learning semantically meaningful features. We show that training U-Nets based on this task is an effective remedy that prevents overfitting and facilitates learning beyond pixel-level features. Shortcut solutions, however, are a big challenge in SSL tasks, including jigsaw puzzles. We propose adversarial robust training as an effective automatic shortcut removal. We achieve competitive or superior results compared to the State of the Art (SOTA) anomaly detection methods on various toy and real-world datasets. Unlike many competitors, the proposed framework is stable, fast, data-efficient, and does not require unprincipled early stopping.
ERL-MPP: Evolutionary Reinforcement Learning with Multi-head Puzzle Perception for Solving Large-scale Jigsaw Puzzles of Eroded Gaps
School of Computer Science, University of Nottingham Ningbo China（宁波诺丁汉大学计算机科学学院）；Nottingham Ningbo China Beacons of Excellence Research and Innovation Institute, University of Nottingham Ningbo China（宁波诺丁汉大学宁波卓越灯塔研究与创新研究所）；Intelligent Modelling & Analysis Group, School of Computer Science, University of Nottingham（诺丁汉大学计算机科学学院智能建模与分析组）；School of Electrical & Electronic Engineering, Nanyang Technological University（南洋理工大学电气与电子工程学院）
Multi-Phase Relaxation Labeling for Square Jigsaw Puzzle Solving
Ben-Gurion University of the Negev（内盖夫本-古里安大学）；Ca’ Foscari University of Venice（威尼斯大学）；Istituto Italiano di Tecnologia（意大利技术研究院）；European Centre of Living Technology（欧洲生命技术中心）
Re-assembling the past: The RePAIR dataset and benchmark for real world 2D and 3D puzzle solving
Fondazione Istituto Italiano di Tecnologia（意大利技术研究院基金会）；Ca’ Foscari University of Venice（威尼斯大学）；Ben-Gurion University of the Negev（本-古里安大学）；Parco Archeologico di Pompei（庞贝考古公园）；University of Lausanne（洛桑大学）；Durham University（杜伦大学）；University of Milan（米兰大学）
Solving Jigsaw Puzzles in the Wild: Human-Guided Reconstruction of Cultural Heritage Fragments
DAIS, Ca’ Foscari University of Venice（威尼斯Ca’ Foscari大学DAIS部门）；Department of Historical Studies, University of Milan（米兰大学历史研究系）；Zhejiang Normal University（浙江师范大学）
VGRP-Bench: Visual Grid Reasoning Puzzle Benchmark for Large Vision-Language Models
School of Computer and Communication Sciences, EPFL（瑞士洛桑联邦理工学院计算机与通信科学学院）；National and Kapodistrian University of Athens（雅典国立卡波迪斯特里亚大学）；Meta GenAI（元宇宙生成式人工智能部门）；University College London（伦敦大学学院）；University of Oxford（牛津大学）
Solving Jigsaw Puzzles By the Graph Connection Laplacian
The University of Arizona（亚利桑那大学）；University of Minnesota, Twin Cities（明尼苏达大学双城分校）；Duke University（杜克大学）；National Center for Theoretical Sciences（国家理论科学中心）
