Object Pose Estimation using Mid-level Visual Representations

Abstract

This work proposes a novel pose estimation model for object categories that can be effectively transferred to previously unseen environments. The deep convolutional network models (CNN) for pose estimation are typically trained and evaluated on datasets specifically curated for object detection, pose estimation, or 3D reconstruction, which requires large amounts of training data. In this work, we propose a model for pose estimation that can be trained with small amount of data and is built on the top of generic mid-level representations [33] (e.g. surface normal estimation and re-shading). These representations are trained on a large dataset without requiring pose and object annotations. Later on, the predictions are refined with a small CNN neural network that exploits object masks and silhouette retrieval. The presented approach achieves superior performance on the Pix3D dataset [26] and shows nearly 35% improvement over the existing models when only 25% of the training data is available. We show that the approach is favorable when it comes to generalization and transfer to novel environments. Towards this end, we introduce a new pose estimation benchmark for commonly encountered furniture categories on challenging Active Vision Dataset [1] and evaluated the models trained on the Pix3D dataset.

Deep Representation Learning: Fundamentals, Technologies, Applications, and Open Challenges

“…Having common latent features acts as a bridge for knowledge transfer. In [450] the authors trained a lightweight CNN module on top of a generic representation called mid-level representation. In comparison to training a complex CNN module which also learns the representations, they achieved superior performance in terms of accuracy, efficiency, and generalization with the method.…”

Modeling the Regional Effects of Climate Change on Future Urban Ozone Air Quality in Tehran, Iran

“…With recent advances in deep learning for pattern recognition, performance of these networks for the task of prediction in different fields of environmental science has progressed even with small amount of training data (Alibak et al, 2022;Nejatishahidin et al, 2022). Application of artificial neural networks (ANN), especially multilayer perceptions (MLP) in the field of air quality has been evaluated in many studies (Chaloulakou et al, 2003;Comrie, 1997;Niska et al, 2004;Schlink et al, 2003;Sousa et al, 2007).…”