Create index elements for the dataset.

Create a synthetic YOLO-format dataset on disk. Generates dummy images with YOLO-format annotations, `data.yaml` file, and directory str

Detect objects in a frame using Inference model, filtering detections by class ID and confidence threshold. Args: frame (np.

Detect objects in a frame using a YOLO model, filtering detections by class ID and confidence threshold. Args: frame (np.nda

Calculates the minimum number of single-character edits required to transform one string into another. Allowed operations are insertion,

Parse results from the Florence 2 multi-model model. https://huggingface.co/microsoft/Florence-2-large Args: result: dict contai

Parse and scale bounding boxes from Google Gemini style [JSON output](https://ai.google.dev/gemini-api/docs/vision?lang=python). The JSO

Parse and scale bounding boxes from moondream JSON output. The JSON is expected to have a key "objects" with a list of dictionaries: {

Parse bounding boxes from paligemma-formatted text, scale them to the specified resolution, and optionally filter by classes. Args:

Create a Color instance from an RGB tuple. Args: color_tuple: A tuple representing the color in RGB format, where each

Creates a Detections instance from [SAM 3](https://github.com/facebookresearch/sam3) inference result. Supports both PVS and

Creates a Detections object from the given result string based on the specified Vision Language Model (VLM). | Name

Get annotation ids that satisfy given filter conditions. Args: img_ids: ids of the images that we want to retrieve.

Get annotations with the specified ids. Args: ids: integer ids specifying annotations. Returns: ann

Get the size category of a bounding boxes array. Args: xyxy: The bounding boxes array shaped (N, 4). Returns: The size

Get image ids that satisfy given filter conditions. Args: img_ids: ids of the images to retrieve. cat_ids: i

Get the public variables of a class instance. Args: instance: The instance of a class include_properties: Whether to include

Retrieves the text labels for the detections. If `custom_labels` are provided, they are used. Otherwise, the labels are extracted from t

Get the size category of detection masks. Args: mask: The mask array shaped (N, H, W), or a :class:`~supervision.detecti

Get the size category of a oriented bounding boxes array. Args: xyxyxyxy: The bounding boxes array shaped (N, 4, 2). Returns:

Calculate the center of a polygon. The center is calculated as the center of the solid figure formed by the points of the polygon Args:

Retrieve the top k class IDs and confidences, ordered in descending order by confidence. Args: k: The number

Compares the data payloads of two Detections instances. Args: data_a, data_b: The data payloads of the instances. Returns:

Converts a binary mask to a list of polygons. Args: mask: A binary mask represented as a 2D NumPy array of shape `(H, W)`,

Wrapper that runs *func* under tracemalloc and returns peak allocation. tracemalloc captures every Python-level allocation — numpy buffers, list

Merges the data payloads of a list of Detections instances. Warning: Assumes that empty detections were filtered-out before passing data to

Args: xyxy: An array of shape `(n, 4)` containing the bounding boxes coordinates in format `[x1, y1, x2, y2]` offset:

Parse a Roboflow API or Inference package result into detection arrays. The returned ``data`` dict always contains ``CLASS_NAME_DATA_FIELD`` as a

Read a json file and return a dict. Args: file_path: The file path as a string or Path object. Returns: A dict of annot

Read a text file and return a list of strings without newline characters. Optionally skip empty lines. Args: file_path: The file

Resize all masks in the array to have a maximum dimension of max_dimension, maintaining aspect ratio. Args: masks: 3D array of b

Converts an RGBA tuple (0-255 each) to a hex color string. Args: rgba: RGBA values in range 0-255. Returns: Hex color s

Shifts `label` bounding boxes into the frame so that they are fully contained within the given resolution, prioritizing the top/left edge.

Returns a list of tracks from track_list_a after removing any tracks that share the same internal_track_id with tracks in track_list_b.

Accumulate per image evaluation results and store the result in self.results

Append one scenario result as a JSON line to *path*. ``math.nan`` (used for skipped dense timings) is serialised as ``null`` so the file is v

Compute the average precision, given the recall and precision curves. Source: https://github.com/rafaelpadilla/Object-Detection-Metri

Calculate insertion anchor in frame to position the center of the count image. Args: line_zone: The line counter object

Find if detections' anchors are within the limit of the line zone and which anchors are on its left and right side. Assumes:

Compute the confusion matrix for each class and IoU threshold. Assumes the matches and prediction_class_ids are sorted by confidence

Compute the confusion matrix for each class and IoU threshold. Assumes the matches and prediction_class_ids are sorted by confidence

Compute the confusion matrix for each class and IoU threshold. Assumes the matches and prediction_class_ids are sorted by confidence

Compute the confusion matrix for each class and IoU threshold. Assumes the matches and prediction_class_ids are sorted by confidence

Compute F1 scores from concatenated stats across all images. ``unique_classes`` is the union of GT and predicted classes so that pred

Compute the IoU between the targets and predictions for a given image and category. Args: img_id: The image id.

Compute precision scores from concatenated stats across all images. ``unique_classes`` is the union of GT and predicted classes so that

Broadcastable function, computing the recall from the confusion matrix. Args: confusion_matrix: shape (N, ..., 3), where

Eigendecompose a 2x2 covariance, returning sorted (eigenvalues, vectors).

Draw the count label on the frame. For example: "out: 7". The label contains horizontal text and is not rotated. Args:

Draw the labels, explaining what each color represents, with automatically computed positions. Args: scene: The

Draw the count label on the frame. For example: "out: 7". The label is oriented to match the line angle. Args: f

Deduplicate matches. If there are multiple matches for the same true or predicted box, only the one with the highest IoU is kept.

Perform evaluation for single category and image. Args: img_id: The image id. cat_id: The category id.

Return two non-overlapping segmentation detections for any tile.

Filter keypoints using a 2D boolean mask of shape `(n, m)`. This method selects the **same set of keypoints from every object**, so e

Assumes that image is in dataset.

Assumes that image is in dataset.