{ "id": 401497, "title": "Learning Materials for Completely New to Structure from Motion", "url": "/competitions/image-matching-challenge-2023/discussion/401497", "author_name": "", "post_date": "2023-04-13T14:39:19.971989900Z", "votes": 75, "comment_count": 6, "views": 0, "content": "

I share with you several nice materials for this competition to obtain basic knowledge.

\n

Spatial Transformation

\n\n

Structure from Motion

\n", "messages": [ { "id": "2220606", "postDate": "04/13/2023 14:39:19", "content": "

I share with you several nice materials for this competition to obtain basic knowledge.

\n

Spatial Transformation

\n\n

Structure from Motion

\n", "rawMarkdown": "I share with you several nice materials for this competition to obtain basic knowledge.\n\n**Spatial Transformation**\n- Computer Graphics (Keenan Crane, Carnegie Mellon University)\n[Lecture 05: Spatial Transformations (CMU 15-462/662)](https://youtu.be/QmFBHSJS0Gw)\n[Lecture 06: 3D Rotations and Complex Representations (CMU 15-462/662)](https://youtu.be/YF5ZUlKxSgE)\n[Lecture 07: Perspective Projection and Texture Mapping (CMU 15-462/662)](https://youtu.be/_4Q4O2Kgdo4)\n\n**Structure from Motion**\n- First Principles of Computer Vision (Shree Nayar, Columbia University)\n[Overview | Structure from Motion](https://youtu.be/oIvg7sbJRIA)\n[Structure from Motion Problem | Structure from Motion](https://youtu.be/JlOzyyhk1v0)\n[Observation Matrix | Structure from Motion](https://youtu.be/Uhkb8Zq-dnM)\n[Rank of Observation Matrix | Structure from Motion](https://youtu.be/Lyd7cf0agvI)\n[Tomasi-Kanade Factorization | Structure from Motion](https://youtu.be/0BVZDyRrYtQ)\n\n- Lecture: Computer Vision (Andreas Geiger, University of Tübingen)\n[Computer Vision - Lecture 3.1 (Structure-from-Motion: Preliminaries)](https://www.youtube.com/watch?v=nrIHi1-a85s&t=912s)\n[Computer Vision - Lecture 3.2 (Structure-from-Motion: Two-frame Structure-from-Motion)](https://www.youtube.com/watch?v=nwTVNpF8SMg)\n[Computer Vision - Lecture 3.3 (Structure-from-Motion: Factorization)](https://www.youtube.com/watch?v=Ti9uotyHOKM)\n[Computer Vision - Lecture 3.4 (Structure-from-Motion: Bundle Adjustment)](https://www.youtube.com/watch?v=MyrVDUnaqUs&t=2s)", "votes": null }, { "id": "2221287", "postDate": "04/14/2023 05:49:12", "content": "

Thank You.

", "rawMarkdown": "Thank You.", "votes": null }, { "id": "2227338", "postDate": "04/19/2023 17:09:30", "content": "

I can imagine someway that you can calculate fundamental matrix, essential matrix or homography, but I cannot think of a way to recover the true scale of translation, as required by the competition. Any ideas?
\nThank you.

", "rawMarkdown": "I can imagine someway that you can calculate fundamental matrix, essential matrix or homography, but I cannot think of a way to recover the true scale of translation, as required by the competition. Any ideas?\nThank you.", "votes": null }, { "id": "2227351", "postDate": "04/19/2023 17:23:41", "content": "

Your reconstruction can be up to scale, please check https://www.kaggle.com/code/eduardtrulls/imc2023-evaluation and in particular the function \"evaluate_R_t\" for further details.

", "rawMarkdown": "Your reconstruction can be up to scale, please check https://www.kaggle.com/code/eduardtrulls/imc2023-evaluation and in particular the function \"evaluate_R_t\" for further details.", "votes": null }, { "id": "2246871", "postDate": "05/05/2023 14:23:13", "content": "

The '3D reconstruction' page on Papers with Code is also a great source for SOTA, papers and code implementation!
\nhttps://paperswithcode.com/task/3d-reconstruction

", "rawMarkdown": "The '3D reconstruction' page on Papers with Code is also a great source for SOTA, papers and code implementation!\n[https://paperswithcode.com/task/3d-reconstruction](url)", "votes": null }, { "id": "2248782", "postDate": "05/07/2023 07:24:46", "content": "

open source methods for 3D Reconstruction with Deep Learning Methods
\nhttps://github.com/natowi/3D-Reconstruction-with-Deep-Learning-Methods

", "rawMarkdown": "open source methods for 3D Reconstruction with Deep Learning Methods\nhttps://github.com/natowi/3D-Reconstruction-with-Deep-Learning-Methods", "votes": null }, { "id": "2276661", "postDate": "05/27/2023 05:32:54", "content": "

It's really helpful. Appreciate for sharing.

", "rawMarkdown": "It's really helpful. Appreciate for sharing.", "votes": null } ], "comments": [ { "id": 2221287, "author_name": "klausmikaelson2002", "author_url": "", "post_date": "04/14/2023 05:49:12", "content": "

Thank You.

", "votes": null, "replies": [] }, { "id": 2227338, "author_name": "ishootlaser", "author_url": "", "post_date": "04/19/2023 17:09:30", "content": "

I can imagine someway that you can calculate fundamental matrix, essential matrix or homography, but I cannot think of a way to recover the true scale of translation, as required by the competition. Any ideas?
\nThank you.

", "votes": null, "replies": [ { "id": 2227351, "author_name": "fabiobellavia", "author_url": "", "post_date": "04/19/2023 17:23:41", "content": "

Your reconstruction can be up to scale, please check https://www.kaggle.com/code/eduardtrulls/imc2023-evaluation and in particular the function \"evaluate_R_t\" for further details.

", "votes": null, "replies": [] } ] }, { "id": 2246871, "author_name": "thomasrochefort", "author_url": "", "post_date": "05/05/2023 14:23:13", "content": "

The '3D reconstruction' page on Papers with Code is also a great source for SOTA, papers and code implementation!
\nhttps://paperswithcode.com/task/3d-reconstruction

", "votes": null, "replies": [] }, { "id": 2248782, "author_name": "raghuhemadri02", "author_url": "", "post_date": "05/07/2023 07:24:46", "content": "

open source methods for 3D Reconstruction with Deep Learning Methods
\nhttps://github.com/natowi/3D-Reconstruction-with-Deep-Learning-Methods

", "votes": null, "replies": [] }, { "id": 2276661, "author_name": "sathvikpeddoju", "author_url": "", "post_date": "05/27/2023 05:32:54", "content": "

It's really helpful. Appreciate for sharing.

", "votes": null, "replies": [] } ], "raw_markdown_by_id": { "2220606": "I share with you several nice materials for this competition to obtain basic knowledge.\n\n**Spatial Transformation**\n- Computer Graphics (Keenan Crane, Carnegie Mellon University)\n[Lecture 05: Spatial Transformations (CMU 15-462/662)](https://youtu.be/QmFBHSJS0Gw)\n[Lecture 06: 3D Rotations and Complex Representations (CMU 15-462/662)](https://youtu.be/YF5ZUlKxSgE)\n[Lecture 07: Perspective Projection and Texture Mapping (CMU 15-462/662)](https://youtu.be/_4Q4O2Kgdo4)\n\n**Structure from Motion**\n- First Principles of Computer Vision (Shree Nayar, Columbia University)\n[Overview | Structure from Motion](https://youtu.be/oIvg7sbJRIA)\n[Structure from Motion Problem | Structure from Motion](https://youtu.be/JlOzyyhk1v0)\n[Observation Matrix | Structure from Motion](https://youtu.be/Uhkb8Zq-dnM)\n[Rank of Observation Matrix | Structure from Motion](https://youtu.be/Lyd7cf0agvI)\n[Tomasi-Kanade Factorization | Structure from Motion](https://youtu.be/0BVZDyRrYtQ)\n\n- Lecture: Computer Vision (Andreas Geiger, University of Tübingen)\n[Computer Vision - Lecture 3.1 (Structure-from-Motion: Preliminaries)](https://www.youtube.com/watch?v=nrIHi1-a85s&t=912s)\n[Computer Vision - Lecture 3.2 (Structure-from-Motion: Two-frame Structure-from-Motion)](https://www.youtube.com/watch?v=nwTVNpF8SMg)\n[Computer Vision - Lecture 3.3 (Structure-from-Motion: Factorization)](https://www.youtube.com/watch?v=Ti9uotyHOKM)\n[Computer Vision - Lecture 3.4 (Structure-from-Motion: Bundle Adjustment)](https://www.youtube.com/watch?v=MyrVDUnaqUs&t=2s)", "2221287": "Thank You.", "2227338": "I can imagine someway that you can calculate fundamental matrix, essential matrix or homography, but I cannot think of a way to recover the true scale of translation, as required by the competition. Any ideas?\nThank you.", "2227351": "Your reconstruction can be up to scale, please check https://www.kaggle.com/code/eduardtrulls/imc2023-evaluation and in particular the function \"evaluate_R_t\" for further details.", "2246871": "The '3D reconstruction' page on Papers with Code is also a great source for SOTA, papers and code implementation!\n[https://paperswithcode.com/task/3d-reconstruction](url)", "2248782": "open source methods for 3D Reconstruction with Deep Learning Methods\nhttps://github.com/natowi/3D-Reconstruction-with-Deep-Learning-Methods", "2276661": "It's really helpful. Appreciate for sharing." }, "source": "meta" }