AdaBins: Depth Estimation using Adaptive Bins

Abstract

We address the problem of estimating a high quality dense depth map from a single RGB input image. We start out with a baseline encoder-decoder convolutional neural network architecture and pose the question of how the global processing of information can help improve overall depth estimation. To this end, we propose a transformer-based architecture block that divides the depth range into bins whose center value is estimated adaptively per image. The final depth values are estimated as linear combinations of the bin centers. We call our new building block AdaBins. Our results show a decisive improvement over the state-of-the-art on several popular depth datasets across all metrics. We also validate the effectiveness of the proposed block with an ablation study and provide the code and corresponding pre-trained weights of the new state-of-the-art model.

13 pages

References in corpus (2)

• PyTorch: An Imperative Style, High-Performance Deep Learning Library
• Depth Map Prediction from a Single Image using a Multi-Scale Deep Network

Cited by in corpus (13)

• MonoViT: Self-Supervised Monocular Depth Estimation with a Vision Transformer
• DepthFormer: Exploiting Long-Range Correlation and Local Information for Accurate Monocular Depth Estimation
• Delving into Deep Imbalanced Regression
• Single Image Depth Estimation: An Overview
• Glass Segmentation with RGB-Thermal Image Pairs
• CARLA-GeAR: a Dataset Generator for a Systematic Evaluation of Adversarial Robustness of Vision Models
• Fast and Accurate Single-Image Depth Estimation on Mobile Devices, Mobile AI 2021 Challenge: Report
• Towards Real-Time Monocular Depth Estimation for Robotics: A Survey
• R4Dyn: Exploring Radar for Self-Supervised Monocular Depth Estimation of Dynamic Scenes
• RadioNet: Transformer based Radio Map Prediction Model For Dense Urban Environments
• Detaching and Boosting: Dual Engine for Scale-Invariant Self-Supervised Monocular Depth Estimation
• A Hybrid mmWave and Camera System for Long-Range Depth Imaging
• Focusing on Shadows for Predicting Heightmaps from Single Remotely Sensed RGB Images with Deep Learning