Kornia 2024: A Year of Growth, Innovation, and Community Engagement

[edgarriba]

Hello Kornia Community!

As we step into a new year, we're proud to share the exciting roadmap for Kornia in 2024. It's been an incredible journey so far, and with over 2 million downloads per month (https://www.pepy.tech/projects/kornia), we're more motivated than ever to continue our growth and innovation in the field of computer vision.

Key Highlights for 2024:

1 General Roadmap: Explore our comprehensive plan for the year at https://docs.google.com/document/d/1zJGDb5Tguri6phR5XVLIgyfR9OGqI-8nQ1kgx4BFEBI/edit?usp=sharing.

Here, we outline the core team's commitments and our dedication to assisting in the development of proposed features that align with our project goals.

2 Release Schedule: Expect consistent updates with one release per quarter, culminating in the v0.8 release by December 2024.

• Q1: Ends on March 31st
• Q2: Ends on June 30th
• Q3: Ends on September 30th
• Q4: Ends on December 31st

3 Infrastructure Changes: We're transitioning from 'master' to 'main' for our first release of the year. This move aims to streamline our processes but may require closing or stalling open PRs older than a few months.

4 Community Support and Sponsorship: We invite the community to support us, especially if older versions of torch are required. Contributions can be made via our GitHub Sponsors profile: https://github.com/sponsors/kornia. We're also updating our BACKERS/SPONSORS documentation and development guidelines.

5 Dynamic Documentation: We're making our documentation more interactive with the addition of GitHub discussions, dynamic plots, and an updated tutorial page. We're also looking into Google Analytics to improve user experience.

New Features and Integrations:

• kornia.optim: Introducing minimum non-linear solvers essential for vision and robotics tasks.
• kornia.tracking: Completing the pyramidal optical flow.
• kornia.sensors: Enhancements in CameraModel, IMU integration, GPS support, and consideration of Lidar for point cloud manipulation.
• Data Structures Refactor: Improving geometry data structures.
• Face and People Detection: Updating algorithms and identifying models for efficient CPU performance.
• Monocular Depth Estimation: Supporting state-of-the-art networks and utilities for depth estimation.
• Deep Dense Stereo Optical Flow: Incorporating advanced networks and utilities for stereo and optical flow.

Get Involved!
Your feedback, suggestions, and contributions are the driving force behind Kornia's success. We encourage you to participate in our community discussions and contribute to the project's growth. Join the community !

Stay tuned for more updates, and let's make 2024 a landmark year for Kornia and the broader computer vision community!

Warm regards,

The Kornia Development Team