Evaluation metrics

The metrics used to evaluate this task will consist on MSE, PSNR and DSSIM [1] which are metrics estimating the distance from the reconstruction to the original. As a specific metric we will use WNJD a metric proposed in this work which evaluates the semantic human knowledge learnt by the model. WNJD works by calculating the performance of a current state of the art method for deep-based pose estimation over the reconstructed images in terms of distances between predicted joint locations and real joint locations. Concrete method will not be explained to participants so that they are not able to tune reconstruction of missing data based on posterior human pose estimation performance.

[1] Z. Wang, A. C. Bovik, H. R. Sheikh, and E. P. Simoncelli, “Image quality assessment: from error visibility to structural similarity,” IEEE transactions on image processing, vol. 13, no. 4, pp. 600–612, 2004.


ECCV Satellite Event and TPAMI

Challenge and associated ECCV Satellite Event and TPAMI special issue on Inpainting and Denoising in the Deep Learning Age!