{"success":true,"data":{"author":"Imageomics Institute (Sun C., WildWing team)","bucketAccessType":"public","bucketName":"imageomics/mmla-pose","bucketPrefix":"","category":"Computer Vision","cloudProvider":"huggingface","createdAt":"2026-11-16T02:38:32.349Z","description":"988 manually labelled 224x224 RGB crops of zebras taken from low-altitude drone footage at the Mpala Research Centre and Ol Pejeta Conservancy in Kenya and The Wilds conservation centre in Ohio, each assigned one of eight pose-orientation classes: front, front-left, left, back-left, back, back-right, right, front-right. The label says which flank faces the camera, which matters for stripe-based individual re-identification and for autonomous drone positioning in the WildWing system (YOLO detector plus a DINOv2 + MLP viewpoint classifier). Classes are imbalanced: 369 right and 295 left side views against 44 to 84 for the oblique and frontal classes; 35 ambiguous crops are kept in a _skip folder. The layout is folder-per-label with filenames encoding source session, DJI video, partition, frame and box index, so each crop traces back to the MMLA and KABR source videos. 1,026 files, about 20 MB; load with torchvision ImageFolder or datasets imagefolder. Dedicated to the public domain under CC0 1.0; cite the dataset and KABR.","displayTitle":"MMLA Pose: Eight-Way Orientation Labels for Zebras in Drone Footage","id":"d9183e43-c05f-5cc5-9de5-bd06cc2a2e92","license":"CC0-1.0","licenseEvidenceUrl":"https://huggingface.co/datasets/imageomics/mmla-pose","objectCountEstimate":1026,"probeStatus":"ok","region":"us","schemaCard":{"accessHints":[{"language":"python-datasets","snippet":"from datasets import load_dataset\nds = load_dataset(\"imageomics/mmla-pose\")"},{"language":"python-hub","snippet":"from huggingface_hub import snapshot_download\nsnapshot_download(repo_id=\"imageomics/mmla-pose\", repo_type=\"dataset\")"}],"caveats":["HuggingFace metadata: author=imageomics, license:cc0-1.0, size_categories:1K<n<10K, format:imagefolder, downloads=547"],"format":"mixed","generatedAt":"2026-11-16T02:38:36.790Z","layout":{"example":"_reference.png","pattern":"{repoId}/{filename}"},"mountPath":"/mnt/archive","sampleHeaders":[{"bytes":"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","encoding":"hex","key":"_reference.png","truncatedAt":512}],"sampleObjects":["_reference.png","_skip/mpala_session_1_DJI_0002_partition_1_DJI_0002_000471_c1_004.jpg","_skip/mpala_session_1_DJI_0002_partition_1_DJI_0002_000921_c0_004.jpg","_skip/mpala_session_1_DJI_0002_partition_1_DJI_0002_001221_c0_004.jpg","_skip/mpala_session_1_DJI_0002_partition_1_DJI_0002_001371_c0_004.jpg"],"sidecars":[{"bytes":"---\nlicense: cc0-1.0\nlanguage:\n- en\npretty_name: \"MMLA Pose: Pose Orientation Labels for Zebras in MMLA Drone Footage\"\ntask_categories:\n- image-classification\ntags:\n- biology\n- image\n- animals\n- CV\n- zebra\n- plains zebra\n- Grevy's zebra\n- pose\n- pose-classification\n- orientation\n- viewpoint\n- drone\n- UAV\n- KABR\n- Mpala Research Centre\n- Ol Pejeta Conservancy\n- Kenya\n- Wilds\nsize_categories:\n- n<1K\ndescription: \"Cropped images of zebras from MMLA drone footage labeled with one of eight pose orientations (front, front-left, left, back-left, back, back-right, right, front-right). Intended as training data for a pose / viewpoint classifier used in autonomous drone navigation around wildlife.\"\n---\n\n# Dataset Card for MMLA Pose: Pose Orientation Labels for Zebras in MMLA Drone Footage\n\nCropped zebra images from [MMLA](https://huggingface.co/collections/imageomics/mmla) drone footage, manually labeled with one of eight discrete pose orientations relative to the camera. The dataset was curated to train a pose / viewpoint classifier (DINOv2 backbone + MLP head) for the [WildWing](https://github.com/Imageomics/wildwing) autonomous drone-navigation system, where knowing which side of an animal is visible drives both individual re-identification (matching against flank stripe patterns) and behavior-aware flight decisions (e.g., avoiding approaches from the front).\n\n|![Pose Classes Reference](https://huggingface.co/datasets/imageomics/mmla-pose/resolve/main/_reference.png)|\n|:--|\n|**Figure 1.** [Pose classes reference diagram](https://huggingface.co/datasets/imageomics/mmla-pose/raw/main/_reference.png). The label indicates which flank of the zebra is visible to the camera, equivalently the direction the zebra's head is pointing relative to the camera. The dark circle in each icon marks the head. The camera/drone is positioned at the bottom of the diagram.|\n\n## Dataset Details\n\n### Dataset Description\n\n- **Curated by:** Claire Sun (curator), with the [Imageomics Institute](https://imageomics.org) and [WildWing](https://imageomics.github.io/wildwing/) project teams\n- **Language(s) (NLP):** en\n- **Homepage:** [MMLA Project](https://imageomics.github.io/mmla/) and [WildWing](https://imageomics.github.io/wildwing/)\n- **Repository:** [WildWing Repo](https://github.com/Imageomics/wildwing)\n- **Related datasets:** [imageomics/mmla_mpala](https://huggingface.co/datasets/imageomics/mmla_mpala), [imageomics/mmla_opc](https://huggingface.co/datasets/imageomics/mmla_opc), and [imageomics/mmla_wilds](https://huggingface.co/datasets/imageomics/mmla_wilds) (source drone footage)\n- **Paper:** N/A (dataset release)\n\nThis dataset contains 988 manually labeled 224x224 RGB crops of zebras observed in aerial drone footage from the [Mpala Research Centre](https://mpala.org/) and [Ol Pejeta Conservancy](https://www.olpejetaconservancy.org/) in Kenya, and at the [Wilds Conservation Center](https://www.thewilds.org/) in Ohio. Each crop is assigned exactly one of eight pose-orientation labels describing which flank of the zebra is visible to the camera (equivalently, the direction of the zebra's head relative to the camera). An additional 35 ambiguous or unusable crops are preserved in a `_skip/` folder to document labeling decisions.\n\nThe labels were created to train the pose component of the [WildWing drone-navigation stack](https://imageomics.github.io/wildwing/). WildWing combines a YOLO-based animal detector with a pose classifier so the drone can reason about how each animal in its view is oriented and make better autonomous decisions about positioning. Reliable pose estimates are particularly important for individual re-identification of zebras, which depends on matching the lateral stripe pattern \\- so the classifier needs to know when a useful side view (left or right profile) is available versus when the animal is facing toward or away from the camera.\n\n### Supported Tasks and Leaderboards\n\n**Image classification (8-class):** Predict the orientation of a zebra in a cropped aerial image. Labels are: `front`, `front-left`, `left`, `back-left`, `back`,","encoding":"utf8","key":"README.md","name":"README.md","size":16404,"truncated":true}],"sourceObjectKey":"_reference.png","starterCell":"from huggingface_hub import hf_hub_download\npath = hf_hub_download(repo_id=\"imageomics/mmla-pose\", filename=\"_reference.png\", repo_type=\"dataset\")\nprint(path)","starterSamplePath":"_reference.png","totalObjects":1026},"summary":"988 manually labelled 224x224 RGB crops of zebras taken from low-altitude drone footage at the Mpala Research Centre and Ol Pejeta Conservancy in Kenya and The Wilds conservation…","title":"MMLA Pose: Eight-Way Orientation Labels for Zebras in Drone Footage","updatedAt":"2026-11-16T16:12:20.378Z","visibility":"public","websiteUrl":"https://imageomics.github.io/wildwing/"}}