Towards Open World Recognition

Abstract

With the of advent rich classification models and high computational power visual recognition systems have found many operational applications. Recognition in the real world poses multiple challenges that are not apparent in controlled lab environments. The datasets are dynamic and novel categories must be continuously detected and then added. At prediction time, a trained system has to deal with myriad unseen categories. Operational systems require minimum down time, even to learn. To handle these operational issues, we present the problem of Open World recognition and formally define it. We prove that thresholding sums of monotonically decreasing functions of distances in linearly transformed feature space can balance "open space risk" and empirical risk. Our theory extends existing algorithms for open world recognition. We present a protocol for evaluation of open world recognition systems. We present the Nearest Non-Outlier (NNO) algorithm which evolves model efficiently, adding object categories incrementally while detecting outliers and managing open space risk. We perform experiments on the ImageNet dataset with 1.2M+ images to validate the effectiveness of our method on large scale visual recognition tasks. NNO consistently yields superior results on open world recognition.

Cited by in corpus (34)

• A continual learning survey: Defying forgetting in classification tasks
• Recent Advances in Open Set Recognition: A Survey
• The Extreme Value Machine
• Class-Incremental Learning: A Survey
• Few-Shot Hyperspectral Image Classification With Unknown Classes Using Multitask Deep Learning
• Open-world Learning and Application to Product Classification
• Collective Decision of One-vs-Rest Networks for Open Set Recognition
• Collective decision for open set recognition
• A Survey on Open Set Recognition
• Denoising Autoencoders for Overgeneralization in Neural Networks
• Unified Probabilistic Deep Continual Learning through Generative Replay and Open Set Recognition
• Are Out-of-Distribution Detection Methods Effective on Large-Scale Datasets?
• Memory Based Online Learning of Deep Representations from Video Streams
• Entropic Out-of-Distribution Detection: Seamless Detection of Unknown Examples
• Open World Object Detection: A Survey
• Into the Unknown: Active Monitoring of Neural Networks
• Vocabulary-informed Zero-shot and Open-set Learning
• Multi-Camera Action Dataset for Cross-Camera Action Recognition Benchmarking
• Open-Set Object Detection Using Classification-free Object Proposal and Instance-level Contrastive Learning
• Entropic Out-of-Distribution Detection
• Boosting Deep Open World Recognition by Clustering
• SMART-Vision: Survey of Modern Action Recognition Techniques in Vision
• Multi-stage Deep Classifier Cascades for Open World Recognition
• Boosting Binary Masks for Multi-Domain Learning through Affine Transformations
• Segmenting Known Objects and Unseen Unknowns without Prior Knowledge
• KS(conf ): A Light-Weight Test if a ConvNet Operates Outside of Its Specifications
• Visual Reasoning: from State to Transformation
• WOT-Class: Weakly Supervised Open-world Text Classification
• Contextual Classification Using Self-Supervised Auxiliary Models for Deep Neural Networks
• torchosr -- a PyTorch extension package for Open Set Recognition models evaluation in Python
• On the Challenges of Open World Recognitionunder Shifting Visual Domains
• Unsupervised detection and open-set classification of fast-ramped flexibility activation events
• Towards Open-World Human Action Segmentation Using Graph Convolutional Networks
• Taking Class Imbalance Into Account in Open Set Recognition Evaluation