Delta-encoder: an effective sample synthesis method for few-shot object recognition

Abstract

Learning to classify new categories based on just one or a few examples is a long-standing challenge in modern computer vision. In this work, we proposes a simple yet effective method for few-shot (and one-shot) object recognition. Our approach is based on a modified auto-encoder, denoted Delta-encoder, that learns to synthesize new samples for an unseen category just by seeing few examples from it. The synthesized samples are then used to train a classifier. The proposed approach learns to both extract transferable intra-class deformations, or "deltas", between same-class pairs of training examples, and to apply those deltas to the few provided examples of a novel class (unseen during training) in order to efficiently synthesize samples from that new class. The proposed method improves over the state-of-the-art in one-shot object-recognition and compares favorably in the few-shot case. Upon acceptance code will be made available.

References in corpus (3)

• Data Augmentation Generative Adversarial Networks
• Few-Example Object Detection with Model Communication
• Predicting Deep Zero-Shot Convolutional Neural Networks using Textual Descriptions

Cited by in corpus (37)

• Generalizing from a Few Examples: A Survey on Few-Shot Learning
• A Comprehensive Survey of Image Augmentation Techniques for Deep Learning
• A Survey on Data Augmentation for Text Classification
• Learning to Self-Train for Semi-Supervised Few-Shot Classification
• Free Lunch for Few-shot Learning: Distribution Calibration
• A Two-Stage Approach to Few-Shot Learning for Image Recognition
• One-Shot Instance Segmentation
• AutoTune: Automatically Tuning Convolutional Neural Networks for Improved Transfer Learning
• Rectifying the Shortcut Learning of Background for Few-Shot Learning
• DeepEMD: Differentiable Earth Mover's Distance for Few-Shot Learning
• Meta-Transfer Learning for Few-Shot Learning
• BriNet: Towards Bridging the Intra-class and Inter-class Gaps in One-Shot Segmentation
• Long-tailed Recognition by Routing Diverse Distribution-Aware Experts
• Adversarial Feature Hallucination Networks for Few-Shot Learning
• Reinforcing Generated Images via Meta-learning for One-Shot Fine-Grained Visual Recognition
• Learning to Focus: Cascaded Feature Matching Network for Few-shot Image Recognition
• A Broader Study of Cross-Domain Few-Shot Learning
• Few-shot Classification via Adaptive Attention
• Hallucination Improves Few-Shot Object Detection
• Rethinking Zero-Shot Learning: A Conditional Visual Classification Perspective
• Dual Variational Generation for Low-Shot Heterogeneous Face Recognition
• Bowtie Networks: Generative Modeling for Joint Few-Shot Recognition and Novel-View Synthesis
• Semantics-driven Attentive Few-shot Learning over Clean and Noisy Samples
• Variational Feature Disentangling for Fine-Grained Few-Shot Classification
• AdarGCN: Adaptive Aggregation GCN for Few-Shot Learning
• MatchingGAN: Matching-based Few-shot Image Generation
• Joint Image-Instance Spatial-Temporal Attention for Few-shot Action Recognition
• Shot in the Dark: Few-Shot Learning with No Base-Class Labels
• Few-shot Learning with Global Relatedness Decoupled-Distillation
• Multi-Agent Semi-Siamese Training for Long-tail and Shallow Face Learning
• Dataset Bias in Few-shot Image Recognition
• Diversity Transfer Network for Few-Shot Learning
• Learning Clusterable Visual Features for Zero-Shot Recognition
• Depth Guided Adaptive Meta-Fusion Network for Few-shot Video Recognition
• Towards Fair Cross-Domain Adaptation via Generative Learning
• Delta-GAN-Encoder: Encoding Semantic Changes for Explicit Image Editing, using Few Synthetic Samples
• Augmented Bi-path Network for Few-shot Learning