A New Distributional Ranking Loss With Uncertainty: Illustrated in Relative Depth Estimation

Mertan, Alican; Sahin, Yusuf Huseyin; Duff, Damien Jade; Ünal, Gözde

doi:10.1109/3dv50981.2020.00118

Cited by 9 publications

(4 citation statements)

References 30 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In addition to the failure cases shown in [61], qualitative results of number of works such as [43] reveals that models fail to estimate the depth correctly for surfaces with figures or reflective surfaces such as mirrors. Intuitively, estimating the depth correctly in these situations requires high-level reasoning, e.g.…”

Section: Discussionmentioning

confidence: 98%

“…An important theme we see here is: Classification and ordinality A number of works decided to formulate the depth estimation problem in a way that does not require the system to estimate the exact depth value [2], [12], [20], [27], [40], [41], [42], [43], [44], [45]. While estimating the relative depth instead of the absolute depth is an alternative, estimating the depth range in a classification setting can be done with success.…”

Section: Fu Et Al [2] Ordinal Regression Ordinal Lossmentioning

confidence: 99%

“…There are two important issues to consider for depth in the wild problem. The first issue a system faces that aims to work in [20] softmax-loss -custom Chen et al [12] pairwise ranking -Hourglass Xian et al [40] pairwise ranking random + hard pairs EncDecResNet Chen et al [41] pairwise ranking random + hard pairs EncDecResNet Xian et al [44] pairwise ranking random + instance edges + image edges EncDecResNet Mertan et al [42] listwise ranking random EncDecResNet Lienen et al [45] listwise ranking random EfficientNet variant Mertan et al [43] distributional ranking random EncDecResNet the wild is that the depth ranges may vary quite a lot which is a problem for the learning process. While the range of absolute depth values for an image of a table may be between 10 cm and a couple of meters, the image of a touristic place such as "Eiffel Tower" contains pixels that are a hundred meters away, if they have a depth value at all (sky regions can be considered infinitely far away).…”

Section: Main Consideration: Working In the Wildmentioning

confidence: 99%

“…In their later work, Mertan et al [43] focus on another aspect of the solutions, which is the confidence of the estimations. They represent the depth of a pixel as a normal distribution, d i ∼ N (µ i , σ 2 i ), and try to minimize the following distributional ranking loss (DL) with respect to d f and d c ,…”

Section: Learning From Relative Depth Annotationsmentioning

confidence: 99%

See 3 more Smart Citations

Single Image Depth Estimation: An Overview

Mertan,

Duff,

Unal

2021

Preprint

Self Cite

View full text Add to dashboard Cite

We review solutions to the problem of depth estimation, arguably the most important subtask in scene understanding. We focus on the single image depth estimation problem. Due to its properties, the single image depth estimation problem is currently best tackled with machine learning methods, most successfully with convolutional neural networks. We provide an overview of the field by examining key works. We examine non-deep learning approaches that mostly predate deep learning and utilize hand-crafted features and assumptions, and more recent works that mostly use deep learning techniques. The single image depth estimation problem is tackled first in a supervised fashion with absolute or relative depth information acquired from human or sensor-labeled data, or in an unsupervised way using unlabelled stereo images or video datasets. We also study multitask approaches that combine the depth estimation problem with related tasks such as semantic segmentation and surface normal estimation. Finally, we discuss investigations into the mechanisms, principles and failure cases of contemporary solutions.

show abstract

Section: Discussionmentioning

confidence: 98%