QtMeshEditor Blazeface is an ONNX-format implementation of Google MediaPipe's BlazeFace short-range face detector. The model is designed to detect faces within images, providing outputs that include bounding box coordinates and six keypoints per anchor. It processes input images of size [1,128,128,3] in RGB format, normalized to the range [-1,1], with aspect ratio preserved through letterboxing using zero borders.
The model outputs two primary tensors: regressors, which provide box and keypoint data for each anchor, and classificators, which yield score logits for detected faces. 3. The detected face region of interest, determined from eye keypoints and expanded to a square region, is intended for further processing by a face mesh landmark model.
0 license. The ONNX graph is a direct conversion of the original Google MediaPipe model, including both code and weights under the same license. The conversion process and its numerical parity with the Python MediaPipe reference implementation are documented and verified through provided scripts and documentation.
evidence_sufficient": true}
QtMeshEditor Blazeface sits in PulseGate's Other AI category. It focuses on providing a fast, open-source face detection model for integration into vision applications. It is built as an open-source project for computer vision developers. QtMeshEditor Blazeface is open source under the MIT license. QtMeshEditor Blazeface is available on the web and API, and it can be self-hosted.
fernandotonon builds and maintains QtMeshEditor Blazeface, and the product first shipped in 2020. Development happens publicly on GitHub with 41 stars and 839 commits in the last 90 days. PulseGate's similarity index finds few close equivalents — QtMeshEditor Blazeface occupies a relatively distinct niche. Key capabilities include face detection, ONNX format, and API integration.
Latest indexed changes and source events
fernandotonon/QtMeshEditor-blazeface-onnx discovered by the PulseGate indexer
Other apps tracked under the same category.
