Learning Single/Multi-Attribute of Object With Symmetry and Group (2021-10-09T00:00:00.000000Z)

TL;DR

A Relative Moving Distance (RMD) based method to utilize the attribute change instead of the attribute pattern itself to classify attributes and is suitable for complex compositions of multiple attributes and objects when incorporating attribute correlations.

Abstract

Attributes and objects can compose diverse compositions. To model the compositional nature of these concepts, it is a good choice to learn them as transformations, e.g., coupling and decoupling. However, complex transformations need to satisfy specific principles to guarantee rationality. Here, we first propose a previously ignored principle of attribute-object transformation: <italic>Symmetry</italic>. For example, coupling <monospace>peeled-apple</monospace> with attribute <monospace>peeled</monospace> should result in <monospace>peeled-apple</monospace>, and decoupling <monospace>peeled</monospace> from <monospace>apple</monospace> should still output <monospace>apple</monospace>. Incorporating the symmetry, we propose a transformation framework inspired by group theory, i.e., SymNet. It consists of two modules: Coupling Network and Decoupling Network. We adopt deep neural networks to implement SymNet and train it in an end-to-end paradigm with the group axioms and symmetry as objectives. Then, we propose a Relative Moving Distance (RMD) based method to utilize the attribute change instead of the attribute pattern itself to classify attributes. Besides the compositions of single-attribute and object, our RMD is also suitable for complex compositions of multiple attributes and objects when incorporating attribute correlations. SymNet can be utilized for attribute learning, compositional zero-shot learning and outperforms the state-of-the-art on four widely-used benchmarks. Code is at <uri>https://github.com/DirtyHarryLYL/SymNet</uri>.

Authors

Cewu Lu

8 papers

Yong-Lu Li

4 papers

Xiaohan Mao

2 papers

TL;DR

Abstract

Authors

References50 items

Learning Graph Embeddings for Compositional Zero-shot Learning

Unsupervised domain adaptation with structural attribute learning networks

HOI Analysis: Integrating and Decomposing Human-Object Interaction

A causal view of compositional zero-shot recognition

Hierarchical Feature Embedding for Attribute Recognition

Detailed 2D-3D Joint Representation for Human-Object Interaction

PaStaNet: Toward Human Activity Knowledge Engine

Symmetry and Group in Attribute-Object Compositions

Improving Pedestrian Attribute Recognition With Weakly-Supervised Multi-Scale Attribute-Specific Localization

Recognizing Unseen Attribute-Object Pair with Generative Model

Unifying Visual Attribute Learning with Object Recognition in a Multiplicative Framework

Task-Driven Modular Networks for Zero-Shot Compositional Learning

TAFE-Net: Task-Aware Feature Embeddings for Low Shot Learning

Learning Compositional Representations for Few-Shot Recognition

Transferable Interactiveness Knowledge for Human-Object Interaction Detection

Learning to Transfer: Generalizable Attribute Learning with Multitask Neural Model Search

Feature Generating Networks for Zero-Shot Learning

Scene Graph Generation from Objects, Phrases and Region Captions

Zero-Shot Learning—A Comprehensive Evaluation of the Good, the Bad and the Ugly

From Red Wine to Red Tomato: Composition with Context

Learning to Detect Human-Object Interactions

Attributes for Improved Attributes: A Multi-Task Network Utilizing Implicit and Explicit Relationships for Facial Attribute Classification

Semantic Jitter: Dense Supervision for Visual Comparisons via Synthetic Images

Fully-Adaptive Feature Sharing in Multi-Task Networks with Applications in Person Attribute Classification

COCO Attributes: Attributes for People, Animals, and Objects

An Empirical Study and Analysis of Generalized Zero-Shot Learning for Object Recognition in the Wild

Learning Attributes Equals Multi-Source Domain Generalization

Visual Genome: Connecting Language and Vision Using Crowdsourced Dense Image Annotations

Deep Residual Learning for Image Recognition

Discovering states and transformations in image collections

Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks

Deep Learning Face Attributes in the Wild

Pedestrian Attribute Recognition At Far Distance

GloVe: Global Vectors for Word Representation

Scene Classification via Hypergraph-Based Semantic Attributes Subnetworks Identification

Inferring Analogous Attributes

Efficient Estimation of Word Representations in Vector Space

UCF101: A Dataset of 101 Human Actions Classes From Videos in The Wild

Relative attributes

A joint learning framework for attribute models and object descriptions

Describing people: A poselet-based approach to attribute classification

Sharing features between objects and their attributes

Caltech-UCSD Birds 200

SUN database: Large-scale scene recognition from abbey to zoo

Describing objects by their attributes

Learning to detect unseen object classes by between-class attribute transfer

ImageNet: A large-scale hierarchical image database

Label-embedding for image classification In PAMI

Visualizing Data using t-SNE

undergraduate student majoring in Computer Science and Engineering

Field of Study

Journal Information

Name

Page

Volume

Venue Information

Name

Type

URL

Alternate Names