3D on KITTI Eigen split unsupervised

Metric: Delta < 1.25^3 (higher is better)

LeaderboardDataset

Loading chart...

Results

Sort:

#	Model↕	Delta < 1.25^3▼	Augmentations	Paper	Date↕	Code
1	Jasmine	0.986	No	Jasmine: Harnessing Diffusion Prior for Self-sup...	2025-03-20	-
2	NimbleD-LiteMono-8M	0.986	No	NimbleD: Enhancing Self-supervised Monocular Dep...	2024-08-26	Code
3	NimbleD-LiteMono	0.986	No	NimbleD: Enhancing Self-supervised Monocular Dep...	2024-08-26	Code
4	Nimbled-SwiftDepth	0.986	No	NimbleD: Enhancing Self-supervised Monocular Dep...	2024-08-26	Code
5	CREMono(M + 1024x320 + Res50)	0.986	No	-	-	-
6	NimbleD-LiteMono-S	0.986	No	NimbleD: Enhancing Self-supervised Monocular Dep...	2024-08-26	Code
7	SPIdepth	0.985	No	SPIdepth: Strengthened Pose Information for Self...	2024-04-18	Code
8	SCIPaD	0.985	No	SCIPaD: Incorporating Spatial Clues into Unsuper...	2024-07-07	Code
9	DCPI-Depth (M+1024x320)	0.985	No	DCPI-Depth: Explicitly Infusing Dense Correspond...	2024-05-27	-
10	MonoViT(MS+1024x320)	0.985	No	MonoViT: Self-Supervised Monocular Depth Estimat...	2022-08-06	Code
11	Manydepth2-NF(M+640x192)	0.985	No	Manydepth2: Motion-Aware Self-Supervised Multi-F...	2023-12-23	Code
12	DCPI-Depth (M+640x192)	0.985	No	DCPI-Depth: Explicitly Infusing Dense Correspond...	2024-05-27	-
13	ProDepth(M+640x192)	0.985	No	ProDepth: Boosting Self-Supervised Multi-Frame M...	2024-07-12	Code
14	Nimbled-MD2-R50	0.985	No	NimbleD: Enhancing Self-supervised Monocular Dep...	2024-08-26	Code
15	Nimbled-SwiftDepth-S	0.985	No	NimbleD: Enhancing Self-supervised Monocular Dep...	2024-08-26	Code
16	Nimbled-MD2-R18	0.985	No	NimbleD: Enhancing Self-supervised Monocular Dep...	2024-08-26	Code
17	PlaneDepth (S + 1280x384)	0.984	No	PlaneDepth: Self-supervised Depth Estimation via...	2022-10-04	Code
18	ProDepth	0.984	No	ProDepth: Boosting Self-Supervised Multi-Frame M...	2024-07-12	Code
19	Manydepth2(M+640x192)	0.984	No	Manydepth2: Motion-Aware Self-Supervised Multi-F...	2023-12-23	Code
20	DIFFNet (MS+1024x320)	0.984	No	Self-Supervised Monocular Depth Estimation with ...	2021-10-18	Code
21	TransDSSL	0.984	No	-	-	Code
22	DS-Depth	0.984	No	DS-Depth: Dynamic and Static Depth Estimation vi...	2023-08-14	Code
23	DynamicDepth (M+640x192)	0.984	No	Disentangling Object Motion and Occlusion for Un...	2022-03-29	Code
24	EPCDepth(S+1024x320)	0.983	No	Excavating the Potential Capacity of Self-Superv...	2021-09-26	Code
25	SCIPaD(M+640x192)	0.983	No	SCIPaD: Incorporating Spatial Clues into Unsuper...	2024-07-07	Code
26	VTDepthB2 (stereo supervision)	0.983	No	-	-	Code
27	X-Distill (M+1024x320)	0.983	No	X-Distill: Improving Self-Supervised Monocular D...	2021-10-24	-
28	VTDepthB2 (monocular supervision)	0.983	No	-	-	Code
29	FeatDepth-MS	0.982	No	Feature-metric Loss for Self-supervised Learning...	2020-07-21	Code
30	EPCDepth(S+640x192)	0.982	No	Excavating the Potential Capacity of Self-Superv...	2021-09-26	Code
31	Dyna-DM	0.982	No	Dyna-DM: Dynamic Object-aware Self-supervised Mo...	2022-06-08	Code
32	SharinGAN	0.981	No	SharinGAN: Combining Synthetic and Real Data for...	2020-06-07	Code

#1Jasmine
0.986
Delta < 1.25^3· 2025-03-20
Jasmine: Harnessing Diffusion Prior for Self-supervised Depth Estimation
#2NimbleD-LiteMono-8MSOTA
0.986
Delta < 1.25^3· 2024-08-26
NimbleD: Enhancing Self-supervised Monocular Depth Estimation with Pseudo-labels and Large-scale Video Pre-training Code
#3NimbleD-LiteMono
0.986
Delta < 1.25^3· 2024-08-26
NimbleD: Enhancing Self-supervised Monocular Depth Estimation with Pseudo-labels and Large-scale Video Pre-training Code
#4Nimbled-SwiftDepth
0.986
Delta < 1.25^3· 2024-08-26
NimbleD: Enhancing Self-supervised Monocular Depth Estimation with Pseudo-labels and Large-scale Video Pre-training Code
#5CREMono(M + 1024x320 + Res50)
0.986
Delta < 1.25^3
No paper
#6NimbleD-LiteMono-S
0.986
Delta < 1.25^3· 2024-08-26
NimbleD: Enhancing Self-supervised Monocular Depth Estimation with Pseudo-labels and Large-scale Video Pre-training Code
#7SPIdepth
0.985
Delta < 1.25^3· 2024-04-18
SPIdepth: Strengthened Pose Information for Self-supervised Monocular Depth Estimation Code
#8SCIPaD
0.985
Delta < 1.25^3· 2024-07-07
SCIPaD: Incorporating Spatial Clues into Unsupervised Pose-Depth Joint Learning Code
#9DCPI-Depth (M+1024x320)
0.985
Delta < 1.25^3· 2024-05-27
DCPI-Depth: Explicitly Infusing Dense Correspondence Prior to Unsupervised Monocular Depth Estimation
#10MonoViT(MS+1024x320)SOTA
0.985
Delta < 1.25^3· 2022-08-06
MonoViT: Self-Supervised Monocular Depth Estimation with a Vision Transformer Code
#11Manydepth2-NF(M+640x192)
0.985
Delta < 1.25^3· 2023-12-23
Manydepth2: Motion-Aware Self-Supervised Multi-Frame Monocular Depth Estimation in Dynamic Scenes Code
#12DCPI-Depth (M+640x192)
0.985
Delta < 1.25^3· 2024-05-27
DCPI-Depth: Explicitly Infusing Dense Correspondence Prior to Unsupervised Monocular Depth Estimation
#13ProDepth(M+640x192)
0.985
Delta < 1.25^3· 2024-07-12
ProDepth: Boosting Self-Supervised Multi-Frame Monocular Depth with Probabilistic Fusion Code
#14Nimbled-MD2-R50
0.985
Delta < 1.25^3· 2024-08-26
NimbleD: Enhancing Self-supervised Monocular Depth Estimation with Pseudo-labels and Large-scale Video Pre-training Code
#15Nimbled-SwiftDepth-S
0.985
Delta < 1.25^3· 2024-08-26
NimbleD: Enhancing Self-supervised Monocular Depth Estimation with Pseudo-labels and Large-scale Video Pre-training Code
#16Nimbled-MD2-R18
0.985
Delta < 1.25^3· 2024-08-26
NimbleD: Enhancing Self-supervised Monocular Depth Estimation with Pseudo-labels and Large-scale Video Pre-training Code
#17PlaneDepth (S + 1280x384)
0.984
Delta < 1.25^3· 2022-10-04
PlaneDepth: Self-supervised Depth Estimation via Orthogonal Planes Code
#18ProDepth
0.984
Delta < 1.25^3· 2024-07-12
ProDepth: Boosting Self-Supervised Multi-Frame Monocular Depth with Probabilistic Fusion Code
#19Manydepth2(M+640x192)
0.984
Delta < 1.25^3· 2023-12-23
Manydepth2: Motion-Aware Self-Supervised Multi-Frame Monocular Depth Estimation in Dynamic Scenes Code
#20DIFFNet (MS+1024x320)SOTA
0.984
Delta < 1.25^3· 2021-10-18
Self-Supervised Monocular Depth Estimation with Internal Feature Fusion Code
#21TransDSSL
0.984
Delta < 1.25^3
No paperCode
#22DS-Depth
0.984
Delta < 1.25^3· 2023-08-14
DS-Depth: Dynamic and Static Depth Estimation via a Fusion Cost Volume Code
#23DynamicDepth (M+640x192)
0.984
Delta < 1.25^3· 2022-03-29
Disentangling Object Motion and Occlusion for Unsupervised Multi-frame Monocular Depth Code
#24EPCDepth(S+1024x320)SOTA
0.983
Delta < 1.25^3· 2021-09-26
Excavating the Potential Capacity of Self-Supervised Monocular Depth Estimation Code
#25SCIPaD(M+640x192)
0.983
Delta < 1.25^3· 2024-07-07
SCIPaD: Incorporating Spatial Clues into Unsupervised Pose-Depth Joint Learning Code
#26VTDepthB2 (stereo supervision)
0.983
Delta < 1.25^3
No paperCode
#27X-Distill (M+1024x320)
0.983
Delta < 1.25^3· 2021-10-24
X-Distill: Improving Self-Supervised Monocular Depth via Cross-Task Distillation
#28VTDepthB2 (monocular supervision)
0.983
Delta < 1.25^3
No paperCode
#29FeatDepth-MSSOTA
0.982
Delta < 1.25^3· 2020-07-21
Feature-metric Loss for Self-supervised Learning of Depth and Egomotion Code
#30EPCDepth(S+640x192)
0.982
Delta < 1.25^3· 2021-09-26
Excavating the Potential Capacity of Self-Supervised Monocular Depth Estimation Code
#31Dyna-DM
0.982
Delta < 1.25^3· 2022-06-08
Dyna-DM: Dynamic Object-aware Self-supervised Monocular Depth Maps Code
#32SharinGANSOTA
0.981
Delta < 1.25^3· 2020-06-07
SharinGAN: Combining Synthetic and Real Data for Unsupervised Geometry Estimation Code