Skip to content

Conversation

@richidubey
Copy link
Collaborator

Merges into #813

## What is a Two-View Estimator?


Two-View Estimator (TVE) takes information about two images and tries to determine their relative pose (how one camera is positioned and oriented with respect to the other one). It also generates correspondences between keypoint in the images.
Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

what information? (feature correspondences)

## What is a Two-View Estimator?


Two-View Estimator (TVE) takes information about two images and tries to determine their relative pose (how one camera is positioned and oriented with respect to the other one). It also generates correspondences between keypoint in the images.
Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

remove: how one camera is positioned and oriented with respect to the other one

use 3D relative orientation and 3D unit translation (or translation direction) to avoid ambiguity.

## What is a Two-View Estimator?


Two-View Estimator (TVE) takes information about two images and tries to determine their relative pose (how one camera is positioned and oriented with respect to the other one). It also generates correspondences between keypoint in the images.
Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

The correspondences are generated in the correspondence_generator, which is outside the two-view estimator. the config mirrors this structure: https://github.com/borglab/gtsfm/blob/master/gtsfm/configs/unified.yaml


Two-View Estimator (TVE) takes information about two images and tries to determine their relative pose (how one camera is positioned and oriented with respect to the other one). It also generates correspondences between keypoint in the images.

As seen in the diagram above, TVE sits between [Correspondence Generator](assets/CORRESPONDENCE_GENERATOR.md) and [Multiview Optimizer](assets/MULTIVIEW_OPTIMIZER.md). TVE does three things in GTSfM:
Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

minor nitpick: this can be an unordered list, remove "3 things" as well. I think the exact number depends on how you break it down, and is not important.

```python
class TwoViewEstimator:

"""Wrapper for running two-view relative pose estimation on image pairs in the dataset."""
Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

please indent the same way we indent code, helps readability.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

3 participants