Attribute Descent: Simulating Object-Centric Datasets on the Content Level and Beyond

TL;DR

This paper introduces Attribute Descent, a method to optimize synthetic data attributes for better alignment with real-world data at the content level, improving training and analysis in object-centric tasks.

Contribution

It presents a novel attribute optimization approach to reduce content-level domain gap in synthetic data for object-centric applications.

Findings

Optimized synthetic data improves classification accuracy.

Synthetic data enhances object re-identification performance.

Method effectively bridges content gap in synthetic datasets.

Abstract

This article aims to use graphic engines to simulate a large number of training data that have free annotations and possibly strongly resemble to real-world data. Between synthetic and real, a two-level domain gap exists, involving content level and appearance level. While the latter is concerned with appearance style, the former problem arises from a different mechanism, i.e, content mismatch in attributes such as camera viewpoint, object placement and lighting conditions. In contrast to the widely-studied appearance-level gap, the content-level discrepancy has not been broadly studied. To address the content-level misalignment, we propose an attribute descent approach that automatically optimizes engine attributes to enable synthetic data to approximate real-world data. We verify our method on object-centric tasks, wherein an object takes up a major portion of an image. In these tasks, the search space is relatively small, and the optimization of each attribute yields sufficiently obvious supervision signals. We collect a new synthetic asset VehicleX, and reformat and reuse existing the synthetic assets ObjectX and PersonX. Extensive experiments on image classification and object re-identification confirm that adapted synthetic data can be effectively used in three scenarios: training with synthetic data only, training data augmentation and numerically understanding dataset content.