Abstract:Optical flow represents vector motion of targets, and its estimation is a hot research in computer vision. The traditional prior-based methods build mathematical models using different regularizers on object characteristics. These methods are straightforward and easily understood, but they suffer from a high computational burden. Deep learning based approaches greatly improve their efficiency, but their filter sizes are fixed and receptive fields are large. To solve these problems, this paper proposes a novel recurrent feature refinement network based on short-term dense connection modules for optical flow estimation (RFRFlow), including the following key parts: firstly, we utilize short-term dense concatenate modules (STDC) as a backbone part for constructing a feature extracting network which can preserve rich details and use global averaging pooling to extract global context information with a large receptive field. Secondly, a 4D correlation volume is constructed using context features of image pairs to extract multi-scale 4D correlation information. Thirdly, a motion feature encoder is reported to transfer the multi-scale 4D correlation information into 2D motion features. Finally, 2D context features and 2D motion features are input into a global motion aggregation (GMA) decoder for pursuing the final optical flow. In the experiments, we firstly present and analyze the results of ablation studies on the selection of batch size and image size. Then, both quantitative and qualitative results on Sintel and KITTI datasets are compared, showing that the proposed RFRFlow is effective and even performs better than the state-of-the-art optical flow estimation methods on performance improvement and objects’ shape preservation.

UDF-GAN: Unsupervised Dense Optical-Flow Estimation Using Cycle Generative Adversarial Networks

Unsupervised Learning of Scene Flow Estimation Fusing with Local Rigidity.

SDOF-GAN: Symmetric Dense Optical Flow Estimation with Generative Adversarial Networks

Learning By Analogy: Reliable Supervision From Transformations For Unsupervised Optical Flow Estimation

Generate Optical Flow with Conditional Generative Adversarial Network

MRDFlow: Unsupervised Optical Flow Estimation Network With Multi-Scale Recurrent Decoder

Unsupervised Deep Learning for Optical Flow Estimation.

SFGAN: Unsupervised Generative Adversarial Learning of 3D Scene Flow from the 3D Scene Self

An Unsupervised Optical Flow Estimation For LiDAR Image Sequences

STFlow: Self-Taught Optical Flow Estimation Using Pseudo Labels

Unsupervised Domain Adaptation for Optical Flow Estimation

NccFlow: Unsupervised Learning of Optical Flow With Non-occlusion from Geometry

Unsupervised Optical Flow Estimation Based on Improved Feature Pyramid

Survey on Unsupervised Learning Methods for Optical Flow Estimation

UnSAMFlow: Unsupervised Optical Flow Guided by Segment Anything Model

Generalized Universal Domain Adaptation with Generative Flow Networks

GCA-Net: A Global Context Aggregation Network for Effective Optical Flow

FG-Depth: Flow-Guided Unsupervised Monocular Depth Estimation

FlowNAS: Neural Architecture Search for Optical Flow Estimation

RFRFlow: Recurrent feature refinement network for optical flow estimation

Learning to Estimate Optical Flow Using Dual-Frequency Paradigm