YORU Your Optimal Recognition Utility

Beta: Video Analysis

Applies to v2.0.0-beta.2. For the stable v1.1.1 procedure, see the User Guides.

Procedure

  1. Select a model to analyze videos.

    Beta 2 accepts YOLOv8 / YOLO11, RT-DETR and torchvision checkpoints, and .onnx files. The backend is chosen from the weights file. YOLOv5 .pt files no longer load — export them to ONNX first, or retrain. See Training.

  2. Select movies.

  3. Select a folder to save results.

  4. Check the previews.

    When a video is loaded, the first video appears in PREVIEW.

    Check for flips, etc., and adjust vertical and horizontal flips if any are present.

  5. Push “YOLO analysis” and start an analysis.

    If you check “Create videos”, YORU will save the videos shown in the box.

    If you check “Tracking algorithm”, YORU will save the IDs in the results csv file.

    YORU has the option of individual identification in multi-animal scenarios, applying the Kuhn-Munkres method (Bashar et al., 2022) to assign IDs based on positional information following object detection. This function is still a beta function.

What changed in Beta 2

  • The window no longer freezes. In Beta 1 the Video Analysis window showed “Not Responding” for the whole job and the progress bar never moved. Analysis now runs on a worker thread with live movie/image progress, remaining time and movies-left counters. Buttons are disabled while busy, and errors are shown in the status line.
  • Rendered videos are no longer upside down. create_video() flipped unconditionally, ignoring the flip checkboxes, and crashed on the last frame. Both are fixed — but check the preview flip settings in step 4, because they now take effect as configured.
  • Detections are no longer lost. A single below-threshold detection used to discard the rest of that frame’s detections in offline analysis.
  • Detection thresholds are uniform across backends (confidence 0.25, IoU 0.45). Previously each backend used its own default, so torchvision models in particular will report a different number of detections than in Beta 1.
  • ONNX inference letterboxes instead of stretching the frame, so boxes on non-square inputs are geometrically correct but numerically different from Beta 1.

Because of the last two points, results produced under Beta 2 are not always directly comparable with results produced under Beta 1 or v1.1.1. Do not mix them within one dataset.

Data example

  • Default


  • Default with tracking


  • Result video frame


Next


Previous