Age | Commit message (Collapse) | Author | |
---|---|---|---|
2022-10-10 | Revert "options/value: add `QObject::connect` wrapper" | Stanislaw Halik | |
This reverts commit a67e8630caf20e7f48151024e9e68dd9271d75c7. | |||
2022-07-01 | options/value: add `QObject::connect` wrapper | Stanislaw Halik | |
This is useful not just to save on complexity in call sites, but also because I plan on using the Verdigris library to remove needless `valueChanged()` and `setValue()` overloads from each `value<t>` instance. Also fix a bug in `options/tie.hpp` where `QComboBox::setCurrentIndex` was erroneously called as `Qt::DirectConnection`. | |||
2022-05-21 | tracker/nn: Fix localizer interfering with tracking when tracking is still good | Michael Welter | |
2022-05-21 | tracker/nn: Fix drawing of uninitialized face ROI buffer | Michael Welter | |
2022-05-19 | tracker/nn: Use postfix underscore to indicate class member variables | Michael Welter | |
2022-05-19 | tracker/nn: Improve preview | Michael Welter | |
* The preview image is now generated with the dimensions of the widget * The pose visualization is added afterwards adjusted to the preview size * The fps / inference time readout is moved to the settings dialog * The actual obtained resolution from the camera is also shown * Dialog layout is changed Note: Switching to using underscore to mark class member vars. It's not consistently applied yet. | |||
2022-05-19 | tracker/nn: Add resolution setting and tooltips | Michael Welter | |
Tooltips only for half of the settings or so. When hovering over the actual input boxes. | |||
2022-05-19 | tracker/nn: Add support for recurrent model and further tweaks | Michael Welter | |
Regarding tweaks: * EWA smoothing of head ROI. Smoothing strength is a UI setting. * Adjustible zoom into the detected face. The predicted ROI is scaled by a factor the user can set. There is a sweet spot somewhere near 1. * Adjustible number of threads * The ROI is no longer taken as model output directly. This was actually not needed. Perhaps as auxiliary training objective for the network. But the tracker implementation now just uses a square area around the head center according to the predicted head size. * Add comment and debug notification on face ROI model output | |||
2022-05-15 | tracker/nn: Support new coordinate conventions for head rotation | Michael Welter | |
2022-05-15 | tracker/nn: Don't hardcode the size of the input image for the pose estimator | Michael Welter | |
2022-05-12 | tracker/nn: catch exception in case the head pose model throws during inference | Michael Welter | |
2022-05-11 | Added support for MJPEG compression for neuralnet | Pedro Silva | |
- Added support for MJPEG compression for neuralnet tracker | |||
2021-12-16 | many: switch from using std::unique_ptr<t>::get() to &*ptr | Stanislaw Halik | |
2021-10-23 | tracker/nn: Remove onnx session options which hurt performance on Linux | Michael Welter | |
2021-10-20 | tracker/nn: Remove unused function | Michael Welter | |
2021-09-26 | compat: use std::clamp, remove own version | Stanislaw Halik | |
2021-09-03 | tracker/neuralnet: fix copying grayscale frameopentrack-2021.1.1 | Stanislaw Halik | |
It crashes later though. | |||
2021-08-29 | tracker/neuralnet: fix compile-time warningopentrack-2021.1.0 | Stanislaw Halik | |
2021-05-29 | TrackerNeuralnet: Display inference time and all available elements | Michael Welter | |
Meaning informative elements like the pose gizmo and bounding boxes | |||
2021-05-29 | NeuralnetTracker: Fix crash due to prediction of tiny bounding box | Michael Welter | |
2021-05-02 | Add a face alignment based tracker | Michael Welter | |
Powered by AI! Models were generated with code from https://github.com/DaWelter/neuralnet-tracker-traincode/releases/tag/v0.1 |