StretchBEV: Stretching Future Instance Prediction
Spatially and Temporally

Adil Kaan Akan Fatma Güney
KUIS AI Center
ECCV 2022

Paper

Suppl. Material

Code

ECCV Video

Poster

Images

Ground Truth

FIERY

StretchBEV-P

Results figure

Results figure

Video

Abstract
In self-driving, predicting future in terms of location and motion of all the agents around the vehicle is a crucial requirement for planning. Recently, a new joint formulation of perception and prediction has emerged by fusing rich sensory information perceived from multiple cameras into a compact bird's-eye view representation to perform prediction. However, the quality of future predictions degrades over time while extending to longer time horizons due to multiple plausible predictions. In this work, we address this inherent uncertainty in future predictions with a stochastic temporal model. Our model learns temporal dynamics in a latent space through stochastic residual updates at each time step. By sampling from a learned distribution at each time step, we obtain more diverse future predictions that are also more accurate compared to previous work, especially stretching both spatially further regions in the scene and temporally over longer time horizons. Despite separate processing of each time step, our model is still efficient through decoupling of the learning of dynamics and the generation of future predictions.
Method Overview
Method overview figure

This figure shows the inference procedure of our model StretchBEV. We start with the first k=3 conditioning frames where we sample the stochastic latent variables from the posterior distribution. On the right, we show the prediction at a step t after the conditioning frames where we sample from the learned future distribution. The dashed vertical line marks the conditioning frames.



Code


Qualitative Examples

Images

Ground Truth

FIERY

StretchBEV-P

Results figure Results figure

Images

Ground Truth

FIERY

StretchBEV-P

Results figure Results figure

Example comparisons with FIERY. From left to right, we show images, ground truth labels, FIERY predictions and StretchBEV-P predictions. We show examples for short (top 2 examples) and mid settings (bottom 2 examples), 2 and 4 seconds into the future respectively.

Sample Comparisons

In this section, we provide additional qualitative examples where we show samples that are generated by FIERY and StretchBEV-P.

FIERY

StretchBEV-P

Results figure Results figure
Results figure Results figure Results figure Results figure

















Paper

StretchBEV: Stretching Future Instance Prediction Spatially and Temporally

Adil Kaan Akan and Fatma Guney

In ECCV, 2022.

@InProceedings{Akan2022ECCV,
            author    = {Akan, Adil Kaan and G\"uney, Fatma},
            title     = {StretchBEV: Stretching Future Instance Prediction Spatially and Temporally},
            journal = {European Conference on Computer Vision (ECCV)},
            year      = {2022},
            }
        


Acknowledgements

Kaan Akan was supported by KUIS AI Center fellowship, Fatma Güney by TUBITAK 2232 International Fellowship for Outstanding Researchers Programme.

The template for this website was borrowed and adapted from Despoina Paschalidou and Niklas Hanselmann.