PK PyTorchKR Papers
Papers with Code · CC BY-SA 4.0

Depth Estimation

3D › Depth Estimation 410 평가 기록 58 datasets

Depth Estimation is the task of measuring the distance of each pixel relative to the camera. Depth is extracted from either monocular (single) or stereo (multiple views of a scene) images. Traditional methods use multi-view geometry to find the relationship between the images. Newer methods can directly estimate depth by minimizing the regression loss, or by learning to generate a novel view from a sequence. The most popular benchmarks are KITTI and NYUv2. Models are typically evaluated according to a RMS metric.

<span class="description-source">Source: DIODE: A Dense Indoor and Outdoor DEpth Dataset </span>

개요는 각 항목의 일부입니다. 위 탭에서 전체 기록을 페이지별로 탐색할 수 있습니다.

벤치마크

Papers with Code 아카이브의 데이터셋별 기록 예시입니다. 평가 조건과 지표 방향이 검증되지 않아 순위를 의미하지 않습니다.

Dataset기록된 모델 예시MetricPaperCode
HybridDepth
RMSE: 0.128absolute relative error: 0.026Delta < 1.25: 0.988Delta < 1.25^2: 1.000
논문 코드
SPIDepth
absolute relative error: 0.029RMSE: 1.394Sq Rel: 0.069RMSE log: 0.048
논문 코드
SPIdepth
absolute relative error: 0.071RMSE: 3.662Sq Rel: 0.531RMSE log: 0.153
논문 코드
HiMODE
RMSE: 0.2619absolute relative error: 0.0532
논문 -
EVP
RMS: 0.224
논문 코드
Distill Any Depth
Delta < 1.25: 0.981absolute relative error: 0.054
논문 코드
Volumetric Propagation Network
RMSE: 636.2
논문 -
IndoorDepth
Root mean square error (RMSE): 0.494Absolute relative error (AbsRel): 0.126delta_1: 84.5delta_2: 96.5
논문 코드
AnyNet
three pixel error: 6.2
논문 코드
FasTCo + OAP
ACER: 21.7
논문 코드
SPIDepth
Abs Rel: 0.299Sq Rel: 1.931RMSE: 6.672
논문 코드
M4Depth+U
Abs Rel: 0.134RMSE log: 0.188
논문 코드

데이터셋

58개
NYU-Depth V2KITTI Eigen splitKITTI Eigen split unsupervisedStanford2D3D PanoramicNYU-Depth V2ETH3DKITTI Depth Completion ValidationNYU-Depth V2 self-supervisedKITTI2015SiW (Protocol 3)Make3DMid-Air Dataset

논문

라이브러리

Papers with Code 리더보드 row에 연결된 구현 저장소입니다.

github.com/AbirKhan96/Intel-ISL-MiDaS 1 논문
github.com/ahmedmostafa0x61/Depth_Estimation 1 논문
github.com/aim-uofa/AdelaiDepth 1 논문
github.com/alexeyab/midas 1 논문
github.com/anlok/depthmap-loktev 1 논문
github.com/freshtan/midas_v2 1 논문
github.com/intel-isl/MiDaS 1 논문
github.com/isl-org/MiDaS 1 논문

하위 태스크

MD Monocular Depth Estimation 25 데이터셋 · 283 rows DA Depth And Camera Motion 8 데이터셋 · 31 rows SD Stereo Depth Estimation 6 데이터셋 · 22 rows SF Stereo-LiDAR Fusion 1 데이터셋 · 9 rows 3D 3D Depth Estimation 2 데이터셋 · 4 rows IM Indoor Monocular Depth Estimation 1 데이터셋 · 2 rows DA Depth Aleatoric Uncertainty Estimation 1 데이터셋 · 1 rows BP Bathymetry prediction 0 데이터셋 · 0 rows DI Depth Image Upsampling 0 데이터셋 · 0 rows DM Depth Map Super-Resolution 0 데이터셋 · 0 rows