Depth Estimation
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.
Papers
Showing 111–120 of 2454 papers
All datasetsStanford2D3D PanoramicNYU-Depth V2DCMeBDthequeScanNetV2Cityscapes testDIODEKITTI 2015Mars DTM EstimationScanNet4D Light Field DatasetKITTI Eigen split
Benchmark Results
| # | Model | Metric | Claimed | Verified | Status |
|---|---|---|---|---|---|
| 1 | OmniDepth | RMSE | 0.62 | — | Unverified |
| 2 | SphereDepth | RMSE | 0.45 | — | Unverified |
| 3 | Jin et al. | RMSE | 0.42 | — | Unverified |
| 4 | BiFuse with fusion | RMSE | 0.41 | — | Unverified |
| 5 | HoHoNet (ResNet-101) | RMSE | 0.38 | — | Unverified |
| 6 | PanoDepth | RMSE | 0.37 | — | Unverified |
| 7 | BiFuse++ | RMSE | 0.37 | — | Unverified |
| 8 | UniFuse with fusion | RMSE | 0.37 | — | Unverified |
| 9 | DisConv | RMSE | 0.37 | — | Unverified |
| 10 | SliceNet | RMSE | 0.37 | — | Unverified |
| # | Model | Metric | Claimed | Verified | Status |
|---|---|---|---|---|---|
| 1 | A2J | mAP | 8.61 | — | Unverified |
| 2 | PAD-Net | RMS | 0.79 | — | Unverified |
| 3 | MS-CRF | RMS | 0.59 | — | Unverified |
| 4 | DORN | RMS | 0.51 | — | Unverified |
| 5 | Freeform | RMS | 0.43 | — | Unverified |
| 6 | Optimized, freeform | RMS | 0.43 | — | Unverified |
| 7 | VNL | RMS | 0.42 | — | Unverified |
| 8 | BTS | RMS | 0.41 | — | Unverified |
| 9 | TransDepth (AGD+ ViT) | RMS | 0.37 | — | Unverified |
| 10 | AdaBins | RMS | 0.36 | — | Unverified |
| # | Model | Metric | Claimed | Verified | Status |
|---|---|---|---|---|---|
| 1 | T2Net | Abs Rel | 0.35 | — | Unverified |
| 2 | MIDAS | Abs Rel | 0.31 | — | Unverified |
| 3 | Bhattacharjee et al. | Abs Rel | 0.25 | — | Unverified |
| # | Model | Metric | Claimed | Verified | Status |
|---|---|---|---|---|---|
| 1 | T2Net | Abs Rel | 0.49 | — | Unverified |
| 2 | MIDAS | Abs Rel | 0.42 | — | Unverified |
| 3 | Bhattacharjee et al. | Abs Rel | 0.38 | — | Unverified |
| # | Model | Metric | Claimed | Verified | Status |
|---|---|---|---|---|---|
| 1 | LeReS | absolute relative error | 0.1 | — | Unverified |
| 2 | DELTAS | absolute relative error | 0.09 | — | Unverified |
| 3 | Distill Any Depth | absolute relative error | 0.04 | — | Unverified |
| # | Model | Metric | Claimed | Verified | Status |
|---|---|---|---|---|---|
| 1 | SDC-Depth | RMSE | 6.92 | — | Unverified |
| 2 | SwinMTL | RMSE | 6.35 | — | Unverified |
| # | Model | Metric | Claimed | Verified | Status |
|---|---|---|---|---|---|
| 1 | H-Net (Ours) | Absolute relative error (AbsRel) | 0.09 | — | Unverified |
| 2 | H-Net (Ours) Full Eigen | Absolute relative error (AbsRel) | 0.08 | — | Unverified |
| # | Model | Metric | Claimed | Verified | Status |
|---|---|---|---|---|---|
| 1 | GLPDepth | Delta < 1.25 | 0.43 | — | Unverified |
| 2 | SRDINET (Model A) | Delta < 1.25 | 0.4 | — | Unverified |
| # | Model | Metric | Claimed | Verified | Status |
|---|---|---|---|---|---|
| 1 | Atlas (finetuned) | RMSE | 0.17 | — | Unverified |
| 2 | Atlas (plain) | RMSE | 0.17 | — | Unverified |
| # | Model | Metric | Claimed | Verified | Status |
|---|---|---|---|---|---|
| 1 | LFattNet | BadPix(0.01) | 17.23 | — | Unverified |
| # | Model | Metric | Claimed | Verified | Status |
|---|---|---|---|---|---|
| 1 | LightDepth | Number of parameters (M) | 42.6 | — | Unverified |
| # | Model | Metric | Claimed | Verified | Status |
|---|---|---|---|---|---|
| 1 | UniFuse | Abs Rel | 0.11 | — | Unverified |
| # | Model | Metric | Claimed | Verified | Status |
|---|---|---|---|---|---|
| 1 | X-TC (Cross-Task Consistency) | L1 error | 1.63 | — | Unverified |