EOPose : Exemplar-based object reposing using Generalized Pose Correspondences

2025-05-06

AI Summary

To address the challenge of rapidly generating high-fidelity product image variants (e.g., novel poses) in e-commerce while preserving color, texture, and brand identity, this paper proposes an unsupervised three-stage pose transfer framework. First, it establishes an exemplar-driven relighting-and-reposing paradigm using generalized pose pairs. Second, it constructs dense correspondences between source and target images via unsupervised keypoint matching. Third, it enables differentiable geometric warping and end-to-end pose-guided rendering. To support training, we curate a paired dataset derived from Objaverse. Experiments demonstrate significant improvements over state-of-the-art baselines in PSNR, SSIM, and FID. User studies and ablation analyses confirm superior photorealism and pose accuracy. To our knowledge, this is the first method enabling general object reposing without annotated keypoints, simultaneously ensuring geometric consistency and texture fidelity.

Technology Category

• Computer Vision: Biometrics, Face, Gesture & Pose
• Machine Learning: Multimodal Learning
• Search and Optimization: Sampling/Simulation-based Search

Application Category

• Economics, Online Markets and Human Computation: Data quality aspects of human-annotated datasets
• Search and Retrieval-Augmented AI: Ad search and search for Web retail
• User Modeling, Personalization and Recommendation: Explainable and interpretable methods for personalization

Abstract

Reposing objects in images has a myriad of applications, especially for e-commerce where several variants of product images need to be produced quickly. In this work, we leverage the recent advances in unsupervised keypoint correspondence detection between different object images of the same class to propose an end-to-end framework for generic object reposing. Our method, EOPose, takes a target pose-guidance image as input and uses its keypoint correspondence with the source object image to warp and re-render the latter into the target pose using a novel three-step approach. Unlike generative approaches, our method also preserves the fine-grained details of the object such as its exact colors, textures, and brand marks. We also prepare a new dataset of paired objects based on the Objaverse dataset to train and test our network. EOPose produces high-quality reposing output as evidenced by different image quality metrics (PSNR, SSIM and FID). Besides a description of the method and the dataset, the paper also includes detailed ablation and user studies to indicate the efficacy of the proposed method

Problem

• Reposing objects in images for e-commerce applications
• Preserving fine-grained details like colors and textures
• Creating high-quality outputs using a novel three-step approach

Innovation

• Uses unsupervised keypoint correspondence detection
• Employs three-step warping and re-rendering approach
• Preserves fine-grained object details accurately