Append detection data to the CSV file. Args: detections: The detection data. custom_data: Custom data to inc

Create an empty Detections object with no bounding boxes, confidences, or class IDs. Returns: An empty Detec

Create a Detections object from list-based inputs. This is a helper function primarily used for testing purposes to quickly instantiate

NumPy-compatible sum with a fast path for per-mask area. When ``axis=(1, 2)``, returns the per-mask True-pixel count via :attr:`area`

Create a :class:`CompactMask` from a dense ``(N, H, W)`` bool array. Bounding boxes are clipped to image bounds and interpreted in the

Returns the color as a BGR tuple. Returns: BGR tuple. Example: ```pycon >>> import supe

Create a Color instance from a hex string. Args: color_hex: The hex string representing the color. This string can

Annotates the given scene with bounding boxes based on the provided detections. Args: scene: The image where bounding bo

Compute the metric from the internal state and return the result.

Materialise all masks as a dense ``(N, H, W)`` boolean array. Returns: Boolean array of shape ``(N, H, W)``. Examples:

Compute Intersection over Union (IoU) of two sets of masks - `masks_true` and `masks_detection`. Accepts both dense ``(N, H, W)`` bo

Return a new CompactMask scaled to a different image resolution. Each crop mask is resized with nearest-neighbour interpolation. Spar

Encode a 2D boolean mask as COCO F-order run lengths (int32 array). Pixels are scanned column-by-column (Fortran order), matching the COCO /

Create a KeyPoints object from list-based inputs. This is a helper function primarily used for testing purposes to quickly instantiate a

Build a CompactMask using full-image bounding boxes (lossless).

Add new predictions and targets to the metric, but do not compute the result. Args: predictions: The predicted detection

Build a CompactMask whose crops equal the full bounding-box extents.

Compute pairwise overlap scores between two sets of oriented bounding boxes using the configured `overlap_metric`. Overlap areas are com

Update the `in_count` and `out_count` based on the objects that cross the line. Args: detections: A Detections object fo

Check whether the `Detections` object has zero bounding boxes. Returns: `True` if there are no detections, `False` other

Run *func* *reps* times and return mean wall-clock seconds per call. When ``parallel > 1``, up to ``parallel`` calls run simultaneously in th

Return the 4 corners of a rotated rectangle as a (4, 2) float32 array.

Converts a 3D `np.array` of 2D bool masks into a 2D `np.array` of bounding boxes. Args: masks: A 3D `np.array` of shape `(N, H, W)`

Decode COCO F-order run lengths back to a 2D boolean mask. This is the shared low-level decoder used by both :func:`rle_to_mask` and :class:`

Assert that the annotated image is mostly the same as the original. Args: original: Original image annotated: Annotated imag

Add new predictions and targets to the metric, but do not compute the result. Args: predictions: The predicted detection

Updates the tracker with the provided detections and returns the updated detection results. Args: detections: Th

Creates and configures a logger with stdout and stderr handlers. This function creates a logger that sends INFO and DEBUG level logs to stdout,

Merge a list of Detections objects into a single Detections object. This method takes a list of Detections objects and combines thei

Return the color at a given index in the palette. Args: idx: Index of the color in the palette. Returns:

Compute pairwise overlap scores between batches of bounding boxes. Supports standard IOU (intersection-over-union) and IOS (intersection

Returns the point as a tuple of integers. Returns: The point as (x, y) integers.

Convert `Detections` to COCO ``annotations`` entries. The internal 0-indexed ``Detections.class_id`` is serialized as a 1-indexed COCO ``cate

Calculates and returns the coordinates of a specific anchor point within the bounding boxes defined by the `xyxy` attribute. The anch

Add new predictions and targets to the metric, but do not compute the result. Args: predictions: The predicted detection

Assert that two values are equal within a specified tolerance. Args: actual: The value to check. expected: The expected valu

Compute pairwise overlap between two :class:`CompactMask` collections. Avoids materialising full ``(N, H, W)`` arrays by: 1. Vectorised boun

Add new predictions and targets to the metric, but do not compute the result. Args: predictions: The predicted detection

Axis-aligned bounding box of a (4, 2) OBB corner array.

Annotates the given scene with masks based on the provided detections. Args: scene: The image where masks will be drawn.

Resize an F-order RLE-encoded crop via nearest-neighbour resampling. Manipulates run lengths directly without decoding to a full 2D boolean a

Calculate Mean Average Precision based on predicted and ground-truth detections at different thresholds using the COCO evaluation met

Calculate confusion matrix based on predicted and ground-truth detections. Args: targets: Detections objects from ground

Create a `sv.Detections` object from the [Roboflow](https://roboflow.com/) API inference result or the [Inference](https://inference.

Processes the current frame, updating time durations for each tracker. Args: detections: The detections for the current frame, in

Draws trace paths on the frame based on the detection coordinates provided. Args: scene: The image on which the traces w

Issue a warning that a function is deprecated. Args: message: The message to display when the function is called.

Writes a single video frame to the target video file. Args: frame: The video frame to be written to the file. The frame

Detections with a CompactMask backed by full-image bounding boxes. Using full-image xyxy means all True pixels are within the crop region, so

Generate *num_masks* random boolean masks with matching tight xyxy boxes. Each mask is built by filling a random sub-rectangle with Bernoulli noi

Decode a single mask crop without allocating the full image array. This is an O(crop_area) operation — ideal for annotators that only

Create an empty KeyPoints object with no key points. Returns: An empty `sv.KeyPoints` object. Examples:

Creates a `sv.Detections` instance from a [YOLOv8](https://github.com/ultralytics/ultralytics) inference result. !!! Note

Calculate the centroids of binary masks in a tensor. Args: masks: A 3D NumPy array of shape (num_masks, height, width).

Download a specified asset if it doesn't already exist or is corrupted. Args: asset_name: The name or type of the asset to be downlo

Determines if elements of a numpy array are present in a list. Args: array (np.ndarray): The numpy array of integers to check. se

Load COCO annotations and convert them to `Detections`. If `force_masks` is `False`, masks are still loaded for images whose annotations

Load polygon zone configurations from a JSON file. This function reads a JSON file which contains polygon coordinates, and converts them

Generate a mask from a polygon. Args: polygon: The polygon for which the mask should be generated, given as a list of vertice

N boxes covering the whole image (ensures crop == full mask).

Build OBB Detections from a list of (4, 2) corner arrays.

Paint each detection's mask into `canvas` in descending-area order. Smaller masks are drawn on top of larger ones. `CompactMask` detections a

Build a DetectionDataset of ``num_images`` 10x10 RGB images on disk, each holding ``dets_per_image`` 1x1 detections of class 0. Image content

Annotates the given scene with labels based on the provided detections. Args: scene: The image where labels will be draw

Convert a KeyPoints object to a Detections object. This approximates the bounding box of the detected object by taking the bo

Convert detections to YOLO annotation lines. Args: detections: The detections to serialize. Each detection must have a valid

Perform Non-Maximum Suppression (NMS) on segmentation predictions. IoU is computed exactly on the full-resolution masks for both dense and

Perform Non-Maximum Suppression on oriented bounding box predictions. Overlap is computed via :func:`oriented_box_iou_batch` on the four

Updates the smoother with a new set of detections from a frame. Args: detections: The detections to add to the smoother.

Return a new :class:`CompactMask` with adjusted offsets and image shape. Used by :class:`~supervision.detection.tools.inference_slicer.Infere

Reference pairwise IoU using the existing dense implementation.

Annotates the scene with a heatmap based on the provided detections. Args: scene: The image where the heatmap will be dr

Exports the dataset to COCO format. This method saves the images and their corresponding annotations in COCO format. !!! tip

Creates a Dataset instance from YOLO formatted data. Args: images_directory_path: The path to the direct

Open the CSV file for writing.

Converts Pillow image into OpenCV image, handling RGB -> BGR conversion. Args: image: Pillow image (in RGB format). Returns

Process video frames asynchronously using a threaded pipeline. This function orchestrates a three-stage pipeline to optimize video processin

Save a DetectionDataset to a COCO-format ``annotations.json`` file. Args: dataset: The DetectionDataset to write. annotation_path

Performs non-max suppression on detection set. Dispatch order: (1) if mask data present, IoU mask is used; (2) else if oriented-box c

Format a speedup/compression ratio with colour coding. ≥10 → green (large win), 1-10 → yellow (modest win), <1 → red (regression). Integer fo

Annotates the given scene by pixelating regions based on the provided detections. Args: scene: The image whe

Annotates the given scene with skeleton vertices based on the provided key points. It draws circles at each key point location. Ancho

Annotates the given scene by drawing lines between specified key points to form edges. Edges where either endpoint is marked as not v

Draw text with background on a scene. Args: scene: A numpy ndarray representing the image, typically of shape (H, W, 3)

Creates a Dataset instance from COCO formatted data. Args: images_directory_path: The path to the direct

Calculate confusion matrix based on predicted and ground-truth detections. Args: predictions: Each element of the list d

Retrieve a subset of the data dictionary based on the given index. Args: data: The data dictionary of the Detections object.