E534 Implementation of Llyod's kmeans clustering algorithm by pmoulon · Pull Request #1476 · openMVG/openMVG · GitHub
[go: up one dir, main page]

Skip to content

Implementation of Llyod's kmeans clustering algorithm#1476

Closed
pmoulon wants to merge 5 commits intoopenMVG:developfrom
rperrot:kmeans
Closed

Implementation of Llyod's kmeans clustering algorithm#1476
pmoulon wants to merge 5 commits intoopenMVG:developfrom
rperrot:kmeans

Conversation

@pmoulon
Copy link
Member
@pmoulon pmoulon commented Feb 11, 2019

No description provided.

@pmoulon pmoulon added this to the 1.5 milestone Feb 11, 2019
@pmoulon pmoulon self-assigned this Feb 11, 2019
pmoulon pushed a commit that referenced this pull request Feb 11, 2019
…#1476

- Fix min/max for std::array
- take @pmoulon remarks into account

- Add kmeans++ initialization (and make it the default init scheme)

- Factorize kmeans code
- Use range loop on one side
- Add full unit test on kmeans
pmoulon added a commit that referenced this pull request Feb 11, 2019
- Style ` ;` -> `;`, ` ,` -> `,`
- use Eigen::Constant for vector initialization
pmoulon added a commit that referenced this pull request Feb 11, 2019
@pmoulon
Copy link
Member Author
pmoulon commented Feb 11, 2019

Pushed in https://github.com/openMVG/openMVG/tree/rperrot-kmeans for further investigation

@pmoulon pmoulon closed this Feb 11, 2019
pmoulon added a commit that referenced this pull request Mar 6, 2019
- Factorize code in kmeans_test
- Remove kmeans_traits dependency to openMVG/matching
- Update kmeans.hpp:
  - use a typed enum for KMeansInitType
  - fix some typo
  - add some more parallelism, and fix a shared variable
  - Use a fix seed to ensure more repeatability
pmoulon added a commit that referenced this pull request Mar 7, 2019
- Fix for Visual Studio (error C3016: 'X': index variable in OpenMP
'for' statement must have signed integral type)
pmoulon added a commit that referenced this pull request Mar 8, 2019
- Add KMeansVectorDataTrait<Eigen::Matrix<T, N, 1>> to replace Vec2,
Vec3, Vec4, Vec2f, Vec3f specialization.
- Use more c++11 (using rather than typedef)
- Use Eigen::Map to enable more SIMD vectorization in ::accumulate and
::divide
pmoulon added a commit that referenced this pull request Mar 13, 2019
- Fix a template specialization (GCC was complaining) Thx to @rperrot
- Fix a cosmetic alignment issue
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants

0