[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"project-9654":3},{"id":4,"name":5,"fullName":6,"owner":5,"repo":5,"description":7,"homepage":8,"htmlUrl":9,"language":10,"languages":9,"totalLinesOfCode":9,"stars":11,"forks":12,"watchers":13,"openIssues":14,"contributorsCount":15,"subscribersCount":15,"size":15,"stars1d":16,"stars7d":17,"stars30d":18,"stars90d":15,"forks30d":15,"starsTrendScore":19,"compositeScore":20,"rankGlobal":9,"rankLanguage":9,"license":21,"archived":22,"fork":22,"defaultBranch":23,"hasWiki":24,"hasPages":22,"topics":25,"createdAt":9,"pushedAt":9,"updatedAt":37,"readmeContent":38,"aiSummary":39,"trendingCount":15,"starSnapshotCount":15,"syncStatus":16,"lastSyncTime":40,"discoverSource":41},9654,"kornia","kornia\u002Fkornia","🐍 Geometric Computer Vision Library for Spatial AI","https:\u002F\u002Fkornia.readthedocs.io",null,"Python",11240,1185,122,34,0,2,13,42,10,83.92,"Apache License 2.0",false,"main",true,[26,27,28,29,30,31,32,33,34,35,36],"artificial-intelligence","computer-vision","deep-learning","hacktoberfest","image-processing","machine-learning","neural-network","python","pytorch","robotics","spatial-ai","2026-06-12 04:00:46","\u003Cdiv align=\"center\">\n\u003Cp align=\"center\">\n  \u003Cimg width=\"55%\" src=\"https:\u002F\u002Fgithub.com\u002Fkornia\u002Fdata\u002Fraw\u002Fmain\u002Fkornia_banner_pixie.png\" \u002F>\n\u003C\u002Fp>\n\n---\n\nEnglish | [简体中文](README_zh-CN.md)\n\n\u003C!-- prettier-ignore -->\n\u003Ca href=\"https:\u002F\u002Fkornia.readthedocs.io\">Docs\u003C\u002Fa> •\n\u003Ca href=\"https:\u002F\u002Fcolab.sandbox.google.com\u002Fgithub\u002Fkornia\u002Ftutorials\u002Fblob\u002Fmaster\u002Fnbs\u002Fhello_world_tutorial.ipynb\">Try it Now\u003C\u002Fa> •\n\u003Ca href=\"https:\u002F\u002Fkornia.github.io\u002Ftutorials\u002F\">Tutorials\u003C\u002Fa> •\n\u003Ca href=\"https:\u002F\u002Fgithub.com\u002Fkornia\u002Fkornia-examples\">Examples\u003C\u002Fa> •\n\u003Ca href=\"https:\u002F\u002Fkornia.github.io\u002F\u002Fkornia-blog\">Blog\u003C\u002Fa> •\n\u003Ca href=\"https:\u002F\u002Fdiscord.gg\u002FHfnywwpBnD\">Community\u003C\u002Fa>\n\n[![PyPI version](https:\u002F\u002Fbadge.fury.io\u002Fpy\u002Fkornia.svg)](https:\u002F\u002Fpypi.org\u002Fproject\u002Fkornia)\n[![Downloads](https:\u002F\u002Fstatic.pepy.tech\u002Fbadge\u002Fkornia)](https:\u002F\u002Fpepy.tech\u002Fproject\u002Fkornia)\n[![star](https:\u002F\u002Fgitcode.com\u002Fkornia\u002Fkornia\u002Fstar\u002Fbadge.svg)](https:\u002F\u002Fgitcode.com\u002Fkornia\u002Fkornia)\n[![Discord](https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FDiscord-5865F2?logo=discord&logoColor=white)](https:\u002F\u002Fdiscord.gg\u002FHfnywwpBnD)\n[![Twitter](https:\u002F\u002Fimg.shields.io\u002Ftwitter\u002Ffollow\u002Fkornia_foss?style=social)](https:\u002F\u002Ftwitter.com\u002Fkornia_foss)\n[![License](https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FLicense-Apache%202.0-blue.svg)](LICENSE)\n\n\u003C\u002Fp>\n\u003C\u002Fdiv>\n\n**Kornia** is a differentiable computer vision library that provides a rich set of differentiable image processing and geometric vision algorithms. Built on top of [PyTorch](https:\u002F\u002Fpytorch.org), Kornia integrates seamlessly into existing AI workflows, allowing you to leverage powerful [batch transformations](), [auto-differentiation]() and [GPU acceleration](). Whether you're working on image transformations, augmentations, or AI-driven image processing, Kornia equips you with the tools you need to bring your ideas to life.\n\n> **📢 Announcement**: Kornia is shifting towards end-to-end vision models. We are focusing on integrating state-of-the-art Vision Language Models (VLM) and Vision Language Agents (VLA) to provide comprehensive end-to-end vision solutions.\n\n## Key Components\n1. **Differentiable Image Processing**\u003Cbr>\n  Kornia provides a comprehensive suite of image processing operators, all differentiable and ready to integrate into deep learning pipelines.\n    - **Filters**: Gaussian, Sobel, Median, Box Blur, etc.\n    - **Transformations**: Affine, Homography, Perspective, etc.\n    - **Enhancements**: Histogram Equalization, CLAHE, Gamma Correction, etc.\n    - **Edge Detection**: Canny, Laplacian, Sobel, etc.\n    - ... check our [docs](https:\u002F\u002Fkornia.readthedocs.io) for more.\n2. **Advanced Augmentations**\u003Cbr>\nPerform powerful data augmentation with Kornia’s built-in functions, ideal for training AI models with complex augmentation pipelines.\n    - **Augmentation Pipeline**: AugmentationSequential, PatchSequential, VideoSequential, etc.\n    - **Automatic Augmentation**: AutoAugment, RandAugment, TrivialAugment.\n3. **AI Models**\u003Cbr>\nLeverage pre-trained AI models optimized for a variety of vision tasks, all within the Kornia ecosystem.\n    - **Face Detection**: YuNet\n    - **Feature Matching**: LoFTR, LightGlue\n    - **Feature Descriptor**: DISK, DeDoDe, SOLD2\n    - **Segmentation**: SAM\n    - **Classification**: MobileViT, VisionTransformer.\n\n\u003Cdetails>\n\u003Csummary>See here for some of the methods that we support! (>500 ops in total !)\u003C\u002Fsummary>\n\n| **Category**               | **Methods\u002FModels**                                                                                                   |\n|----------------------------|---------------------------------------------------------------------------------------------------------------------|\n| **Image Processing**        | - Color conversions (RGB, Grayscale, HSV, etc.)\u003Cbr>- Geometric transformations (Affine, Homography, Resizing, etc.)\u003Cbr>- Filtering (Gaussian blur, Median blur, etc.)\u003Cbr>- Edge detection (Sobel, Canny, etc.)\u003Cbr>- Morphological operations (Erosion, Dilation, etc.)                                 |\n| **Augmentation**            | - Random cropping, Erasing\u003Cbr> - Random geometric transformations (Affine, flipping, Fish Eye, Perspecive, Thin plate spline, Elastic)\u003Cbr>- Random noises (Gaussian, Median, Motion, Box, Rain, Snow, Salt and Pepper)\u003Cbr>- Random color jittering (Contrast, Brightness, CLAHE, Equalize, Gamma, Hue, Invert, JPEG, Plasma, Posterize, Saturation, Sharpness, Solarize)\u003Cbr> - Random MixUp, CutMix, Mosaic, Transplantation, etc.                  |\n| **Feature Detection**       | - Detector (Harris, GFTT, Hessian, DoG, KeyNet, DISK and DeDoDe)\u003Cbr> - Descriptor (SIFT, HardNet, TFeat, HyNet, SOSNet, and LAFDescriptor)\u003Cbr>- Matching (nearest neighbor, mutual nearest neighbor, geometrically aware matching, AdaLAM LightGlue, and LoFTR)                    |\n| **Geometry**                | - Camera models and calibration\u003Cbr>- Stereo vision (epipolar geometry, disparity, etc.)\u003Cbr>- Homography estimation\u003Cbr>- Depth estimation from disparity\u003Cbr>- 3D transformations                |\n| **Deep Learning Layers**    | - Custom convolution layers\u003Cbr>- Recurrent layers for vision tasks\u003Cbr>- Loss functions (e.g., SSIM, PSNR, etc.)\u003Cbr>- Vision-specific optimizers                                        |\n| **Photometric Functions**   | - Photometric loss functions\u003Cbr>- Photometric augmentations                                                                                           |\n| **Filtering**               | - Bilateral filtering\u003Cbr>- DexiNed\u003Cbr>- Dissolving\u003Cbr>- Guided Blur\u003Cbr>- Laplacian\u003Cbr>- Gaussian\u003Cbr>- Non-local means\u003Cbr>- Sobel\u003Cbr>- Unsharp masking                                                                                            |\n| **Color**                   | - Color space conversions\u003Cbr>- Brightness\u002Fcontrast adjustment\u003Cbr>- Gamma correction                                                                       |\n| **Stereo Vision**           | - Disparity estimation\u003Cbr>- Depth estimation\u003Cbr>- Rectification                                                                                           |\n| **Image Registration**      | - Affine and homography-based registration\u003Cbr>- Image alignment using feature matching                                                                     |\n| **Pose Estimation**         | - Essential and Fundamental matrix estimation\u003Cbr>- PnP problem solvers\u003Cbr>- Pose refinement                                                                |\n| **Optical Flow**            | - Farneback optical flow\u003Cbr>- Dense optical flow\u003Cbr>- Sparse optical flow                                                                                  |\n| **3D Vision**               | - Depth estimation\u003Cbr>- Point cloud operations\u003Cbr>                                                                |\n| **Image Denoising**         | - Gaussian noise removal\u003Cbr>- Poisson noise removal                                                                                                        |\n| **Edge Detection**          | - Sobel operator\u003Cbr>- Canny edge detection                                                                                                                 |                                               |\n| **Transformations**         | - Rotation\u003Cbr>- Translation\u003Cbr>- Scaling\u003Cbr>- Shearing                                                                                                     |\n| **Loss Functions**          | - SSIM (Structural Similarity Index Measure)\u003Cbr>- PSNR (Peak Signal-to-Noise Ratio)\u003Cbr>- Cauchy\u003Cbr>- Charbonnier\u003Cbr>- Depth Smooth\u003Cbr>- Dice\u003Cbr>- Hausdorff\u003Cbr>- Tversky\u003Cbr>- Welsch\u003Cbr>                                   |                                                                                             |\n| **Morphological Operations**| - Dilation\u003Cbr>- Erosion\u003Cbr>- Opening\u003Cbr>- Closing                                                                                                          |\n\n\u003C\u002Fdetails>\n\n## Half-Precision Support\n\n| Module | float16 | bfloat16 | Notes |\n|--------|:-------:|:--------:|-------|\n| `kornia.color` | ⚠️ | ⚠️ | Most conversions work for both; FFT-based ops may fail |\n| `kornia.filters` | ⚠️ | ⚠️ | Basic filters work; FFT-based ops may fail on CUDA |\n| `kornia.enhance` | ⚠️ | ⚠️ | Histogram eq \u002F gamma \u002F ZCA work (linalg ops use cast helpers) |\n| `kornia.morphology` | ✅ | ✅ | Pure conv\u002Fpool ops; no dtype restrictions |\n| `kornia.augmentation` | ⚠️ | ⚠️ | Most ops work; precision-sensitive transforms may be inaccurate |\n| `kornia.geometry.transform` | ⚠️ | ⚠️ | Affine\u002Fwarp\u002Fresize work via cast helpers; thin-plate spline may fail |\n| `kornia.geometry.camera` | ⚠️ | ⚠️ | Pinhole model and most camera ops work; `StereoCamera` accepts both |\n| `kornia.geometry.calibration` | ❌ | ❌ | Explicitly accepts float32\u002Ffloat64 only (PnP solver) |\n| `kornia.geometry.epipolar` | ⚠️ | ⚠️ | SVD\u002Finverse use cast helpers; both dtypes work |\n| `kornia.geometry.homography` | ⚠️ | ⚠️ | Uses `_torch_svd_cast` — both dtypes work via casting |\n| `kornia.geometry.liegroup` | ⚠️ | ⚠️ | Most ops work via cast helpers; some linalg paths may fail |\n| `kornia.geometry.solvers` | ⚠️ | ⚠️ | Uses `_torch_solve_cast` — both dtypes work via casting |\n| `kornia.geometry.subpix` | ⚠️ | ⚠️ | Soft-argmax works; precision-sensitive ops may be inaccurate |\n| `kornia.losses` | ⚠️ | ⚠️ | Photometric losses work; linalg-based losses may not |\n| `kornia.feature` | ⚠️ | ⚠️ | Detectors\u002Fdescriptors work; matching uses manual cdist fallback |\n| `kornia.metrics` | ⚠️ | ⚠️ | Pixel-level metrics work; linalg-based metrics may not |\n| `kornia.models` | ⚠️ | ⚠️ | Conv-based models work; attention-based models may have dtype mismatches |\n\n✅ Supported   ⚠️ Partial   ❌ Not supported\n\n**Test results** (commit `6131e98`, 2026-03-21):\n\n| Run | Passed | Failed | Skipped | Pass% |\n|-----|-------:|-------:|--------:|------:|\n| CPU float32 *(baseline)* | 7647 | 3 | 3269 | **99.9%** |\n| CUDA float32 *(baseline)* | 7634 | 3 | 3280 | **99.9%** |\n| CPU float16 | 6866 | 747 | 3306 | **90.1%** |\n| CPU bfloat16 | 6838 | 812 | 3269 | **89.3%** |\n| CUDA float16 *(KORNIA_TEST_IN_SUBPROCESS=1)* | 6727 | 643 | 3556 | **91.3%** |\n| CUDA bfloat16 *(KORNIA_TEST_IN_SUBPROCESS=1)* | 6695 | 713 | 3518 | **90.4%** |\n\nSee the [full precision guide](https:\u002F\u002Fkornia.readthedocs.io\u002Fen\u002Fstable\u002Fget-started\u002Fprecision.html) for details.\n\n## Sponsorship\n\nKornia is an open-source project that is developed and maintained by volunteers. Whether you're using it for research or commercial purposes, consider sponsoring or collaborating with us. Your support will help ensure Kornia's growth and ongoing innovation. Reach out to us today and be a part of shaping the future of this exciting initiative!\n\n\u003Ca href=\"https:\u002F\u002Fopencollective.com\u002Fkornia\u002Fdonate\" target=\"_blank\">\n  \u003Cimg src=\"https:\u002F\u002Fopencollective.com\u002Fwebpack\u002Fdonate\u002Fbutton@2x.png?color=blue\" width=300 \u002F>\n\u003C\u002Fa>\n\n## Installation\n\n[![PyPI python](https:\u002F\u002Fimg.shields.io\u002Fpypi\u002Fpyversions\u002Fkornia)](https:\u002F\u002Fpypi.org\u002Fproject\u002Fkornia)\n[![pytorch](https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FPyTorch_2.0.0+-ee4c2c?logo=pytorch&logoColor=white)](https:\u002F\u002Fpytorch.org\u002Fget-started\u002Flocally\u002F)\n\n### From pip\n\n  ```bash\n  pip install kornia\n  ```\n\n\u003Cdetails>\n  \u003Csummary>Other installation options\u003C\u002Fsummary>\n\n#### From source with editable mode\n\n  ```bash\n  pip install -e .\n  ```\n\n#### For development with Pixi (Recommended)\n\nFor development, Kornia uses [pixi](https:\u002F\u002Fpixi.sh) for fast Python package management and environment management. The project includes a `pixi.toml` configuration file for reproducible dependency management.\n\n  ```bash\n  # Install pixi (if not already installed)\n  curl -fsSL https:\u002F\u002Fpixi.sh\u002Finstall.sh | bash\n\n  # Install dependencies and set up the development environment\n  pixi install\n\n  # Run tests\n  pixi run test\n\n  # For CUDA development\n  pixi run -e cuda install\n  pixi run -e cuda test-cuda\n  ```\n\nThis will set up a complete development environment with all dependencies. For more details on dependency management and available tasks, see [CONTRIBUTING.md](CONTRIBUTING.md).\n\n#### From Github url (latest version)\n\n  ```bash\n  pip install git+https:\u002F\u002Fgithub.com\u002Fkornia\u002Fkornia\n  ```\n\n\u003C\u002Fdetails>\n\n## Quick Start\n\nKornia is not just another computer vision library — it's your gateway to effortless Computer Vision and AI.\n\n\u003Cdetails>\n\u003Csummary>Get started with Kornia image transformation and augmentation!\u003C\u002Fsummary>\n\n```python\nimport numpy as np\nimport kornia_rs as kr\n\nfrom kornia.augmentation import AugmentationSequential, RandomAffine, RandomBrightness\nfrom kornia.filters import StableDiffusionDissolving\n\n# Load and prepare your image\nimg: np.ndarray = kr.read_image_any(\"img.jpeg\")\nimg = kr.resize(img, (256, 256), interpolation=\"bilinear\")\n\n# alternatively, load image with PIL\n# img = Image.open(\"img.jpeg\").resize((256, 256))\n# img = np.array(img)\n\nimg = np.stack([img] * 2)  # batch images\n\n# Define an augmentation pipeline\naugmentation_pipeline = AugmentationSequential(\n    RandomAffine((-45., 45.), p=1.),\n    RandomBrightness((0.,1.), p=1.)\n)\n\n# Leveraging StableDiffusion models\ndslv_op = StableDiffusionDissolving()\n\nimg = augmentation_pipeline(img)\ndslv_op(img, step_number=500)\n\ndslv_op.save(\"Kornia-enhanced.jpg\")\n```\n\n\u003C\u002Fdetails>\n\n\u003Cdetails>\n\u003Csummary>Find out Kornia ONNX models with ONNXSequential!\u003C\u002Fsummary>\n\n```python\nimport numpy as np\nfrom kornia.onnx import ONNXSequential\n# Chain ONNX models from HuggingFace repo and your own local model together\nonnx_seq = ONNXSequential(\n    \"hf:\u002F\u002Foperators\u002Fkornia.geometry.transform.flips.Hflip\",\n    \"hf:\u002F\u002Fmodels\u002Fkornia.models.detection.rtdetr_r18vd_640x640\",  # Or you may use \"YOUR_OWN_MODEL.onnx\"\n)\n# Prepare some input data\ninput_data = np.random.randn(1, 3, 384, 512).astype(np.float32)\n# Perform inference\noutputs = onnx_seq(input_data)\n# Print the model outputs\nprint(outputs)\n\n# Export a new ONNX model that chains up all three models together!\nonnx_seq.export(\"chained_model.onnx\")\n```\n\u003C\u002Fdetails>\n\n## Multi-framework support\n\nYou can now use Kornia with [TensorFlow](https:\u002F\u002Fwww.tensorflow.org\u002F), [JAX](https:\u002F\u002Fjax.readthedocs.io\u002Fen\u002Flatest\u002Findex.html), and [NumPy](https:\u002F\u002Fnumpy.org\u002F). See [Multi-Framework Support](docs\u002Fsource\u002Fget-started\u002Fmulti-framework-support.rst) for more details.\n\n```python\nimport kornia\ntf_kornia = kornia.to_tensorflow()\n```\n\n\u003Cp align=\"center\">\n  Powered by\n  \u003Ca href=\"https:\u002F\u002Fgithub.com\u002Fivy-llc\u002Fivy\" target=\"_blank\">\n    \u003Cdiv class=\"dark-light\" style=\"display: block;\" align=\"center\">\n      \u003Cimg class=\"dark-light\" width=\"15%\" src=\"https:\u002F\u002Fraw.githubusercontent.com\u002Fivy-llc\u002Fassets\u002Frefs\u002Fheads\u002Fmain\u002Fassets\u002Flogos\u002Fivy-long.svg\"\u002F>\n    \u003C\u002Fdiv>\n  \u003C\u002Fa>\n\u003C\u002Fp>\n\n## Call For Contributors\n\nAre you passionate about computer vision, AI, and open-source development? Join us in shaping the future of Kornia! We are actively seeking contributors to help expand and enhance our library, making it even more powerful, accessible, and versatile. Whether you're an experienced developer or just starting, there's a place for you in our community.\n\n### Accessible AI Models\n\nWe are excited to announce our latest advancement: a new initiative designed to seamlessly integrate lightweight AI models into Kornia.\nWe aim to run any models as smooth as big models such as StableDiffusion, to support them well in many perspectives.\n\n**Priority Focus: VLM\u002FVLA Models**\n\nOur primary focus is on integrating **Vision Language Models (VLM)** and **Vision Language Agents (VLA)** to enable end-to-end vision solutions. We're actively seeking contributors to help us:\n\n- **VLM\u002FVLA Integration (Priority)**: Implement and integrate state-of-the-art Vision Language Models and Vision Language Agents. This includes models like Qwen2.5-VL, SAM-3, and other cutting-edge VLM\u002FVLA architectures. If you are a researcher working on VLM\u002FVLA models, Kornia is an excellent place for you to promote your model!\n- Expand the Model Selection: Import decent models into our library. If you are a researcher, Kornia is an excellent place for you to promote your model!\n- Model Optimization: Work on optimizing models to reduce their computational footprint while maintaining accuracy and performance. You may start from offering ONNX support!\n- Model Documentation: Create detailed guides and examples to help users get the most out of these models in their projects.\n\n### Documentation And Tutorial Optimization\n\nKornia's foundation lies in its extensive collection of classic computer vision operators, providing robust tools for image processing, feature extraction, and geometric transformations. We continuously seek for contributors to help us improve our documentation and present nice tutorials to our users.\n\n\n## Cite\n\nIf you are using kornia in your research-related documents, it is recommended that you cite the paper. See more in [CITATION](.\u002FCITATION.md).\n\n  ```bibtex\n  @inproceedings{eriba2019kornia,\n    author    = {E. Riba, D. Mishkin, D. Ponsa, E. Rublee and G. Bradski},\n    title     = {Kornia: an Open Source Differentiable Computer Vision Library for PyTorch},\n    booktitle = {Winter Conference on Applications of Computer Vision},\n    year      = {2020},\n    url       = {https:\u002F\u002Farxiv.org\u002Fpdf\u002F1910.02190.pdf}\n  }\n  ```\n\n## Contributing\n\nWe appreciate all contributions. If you are planning to contribute back bug-fixes, please do so without any further discussion. If you plan to contribute new features, utility functions or extensions, please first open an issue and discuss the feature with us. Please, consider reading the [CONTRIBUTING](.\u002FCONTRIBUTING.md) notes. The participation in this open source project is subject to [Code of Conduct](.\u002FCODE_OF_CONDUCT.md).\n\n### AI Policy\n\nKornia accepts AI-assisted code but strictly rejects AI-generated contributions where the submitter acts as a proxy. All contributors must be the **Sole Responsible Author** for every line of code. Please review our [AI Policy](AI_POLICY.md) before submitting pull requests. Key requirements include:\n\n- **Proof of Verification**: PRs must include local test logs proving execution\n- **Pre-Discussion**: All PRs must be discussed in Discord or via a GitHub issue before implementation\n- **Library References**: Implementations must be based on existing library references (PyTorch, OpenCV, etc.)\n- **Use Existing Utilities**: Use existing `kornia` utilities instead of reinventing the wheel\n- **Explain It**: You must be able to explain any code you submit\n\nAutomated AI reviewers (e.g., GitHub Copilot) will check PRs against these policies. See [AI_POLICY.md](AI_POLICY.md) for complete details.\n\n## Community\n- **Discord:** Join our workspace to keep in touch with our core contributors, get latest updates on the industry and  be part of our community. [JOIN HERE](https:\u002F\u002Fdiscord.gg\u002FHfnywwpBnD)\n- **GitHub Issues:** bug reports, feature requests, install issues, RFCs, thoughts, etc. [OPEN](https:\u002F\u002Fgithub.com\u002Fkornia\u002Fkornia\u002Fissues\u002Fnew\u002Fchoose)\n- **Forums:** discuss implementations, research, etc. [GitHub Forums](https:\u002F\u002Fgithub.com\u002Fkornia\u002Fkornia\u002Fdiscussions)\n\n\u003Ca href=\"https:\u002F\u002Fgithub.com\u002FKornia\u002Fkornia\u002Fgraphs\u002Fcontributors\">\n  \u003Cimg src=\"https:\u002F\u002Fcontrib.rocks\u002Fimage?repo=Kornia\u002Fkornia\" width=\"60%\" \u002F>\n\u003C\u002Fa>\n\nMade with [contrib.rocks](https:\u002F\u002Fcontrib.rocks).\n\n## License\n\nKornia is released under the Apache 2.0 license. See the [LICENSE](.\u002FLICENSE) file for more information.\n","Kornia 是一个基于 PyTorch 的可微分计算机视觉库，提供了丰富的可微分图像处理和几何视觉算法。其核心功能包括多种图像处理算子（如滤波、变换、增强和边缘检测）、强大的数据增强工具（如自动增强和序列化增强）以及预训练的 AI 模型，适用于各种视觉任务。Kornia 通过无缝集成到现有的 AI 工作流中，支持批处理转换、自动微分和 GPU 加速。该库特别适合需要进行图像变换、数据增强或 AI 驱动图像处理的应用场景，例如自动驾驶、机器人视觉和空间感知等。","2026-06-11 03:24:00","top_topic"]