View on GitHub

Eliminating the Blind Spot: Adapting 3D Object Detection and Monocular Depth Estimation to 360° Panoramic Imagery

Grégoire Payen de La Garanderie, Amir Atapour Abarghouei, Toby P. Breckon

About

Recent automotive vision work has focused almost exclusively on processing forward-facing cameras. However, future autonomous vehicles will not be viable without a more comprehensive surround sensing, akin to a human driver, as can be provided by 360° panoramic cameras. We present an approach to adapt contemporary deep network architectures developed on conventional rectilinear imagery to work on equirectangular 360° panoramic imagery. To address the lack of annotated panoramic automotive datasets availability, we adapt contemporary automotive dataset, via style and projection transformations, to facilitate the cross-domain retraining of contemporary algorithms for panoramic imagery. Following this approach we retrain and adapt existing architectures to recover scene depth and 3D pose of vehicles from monocular panoramic imagery without any panoramic training labels or calibration parameters. Our approach is evaluated qualitatively on crowd-sourced panoramic images and quantitatively using an automotive environment simulator to provide the first benchmark for such techniques within panoramic imagery.

Publication

Eliminating the Blind Spot: Adapting 3D Object Detection and Monocular Depth Estimation to 360° Panoramic Imagery (G. Payen de La Garanderie, A. Atapour Abarghouei, T.P. Breckon), In Proc. European Conference on Computer Vision, Springer, 2018. [arxiv link][ECCV link].

Code and Dataset

Our synthetic dataset of images generated using the Carla simulator is available here.

Note that the dataset was updated on the 25/02/2020 to improve the ground truth bounding box quality and add 3D object detection evaluation metrics. The original dataset is still available here.

The object detection code is available on GitHub. The monocular depth estimation code is available on Github.