MoNet3D: Towards Accurate Monocular 3D Object Localization in Real Time

TL;DR

MoNet3D introduces a novel deep learning framework that accurately localizes objects in 3D space from monocular images in real time, leveraging spatial geometric priors to enhance precision and speed.

Contribution

The paper presents MoNet3D, a new method integrating spatial geometric priors into neural networks for monocular 3D object localization, achieving high accuracy and real-time performance.

Findings

Depth prediction accuracy of 96.25%

Horizontal coordinate accuracy of 94.74%

Real-time processing at 27.85 FPS

Abstract

Monocular multi-object detection and localization in 3D space has been proven to be a challenging task. The MoNet3D algorithm is a novel and effective framework that can predict the 3D position of each object in a monocular image and draw a 3D bounding box for each object. The MoNet3D method incorporates prior knowledge of the spatial geometric correlation of neighbouring objects into the deep neural network training process to improve the accuracy of 3D object localization. Experiments on the KITTI dataset show that the accuracy for predicting the depth and horizontal coordinates of objects in 3D space can reach 96.25\% and 94.74\%, respectively. Moreover, the method can realize the real-time image processing at 27.85 FPS, showing promising potential for embedded advanced driving-assistance system applications. Our code is publicly available at https://github.com/CQUlearningsystemgroup/YicongPeng.