Ultra-lightweight human detection. The number of parameters does not correlate to inference speed. For limited use cases, an input image resolution of 64x64 is sufficient. High-level object detecti…

This is a hand gesture recognition program that replaces the entire MediaPipe process with ONNX. Simultaneous detection of multiple palms and a simple tracker are additionally implemented. In addit…

Self-Created Tools to convert ONNX files (NCHW) to TensorFlow/TFLite/Keras format (NHWC). The purpose of this tool is to solve the massive Transpose extrapolation problem in onnx-tensorflow (onnx-t…

Ultra-lightweight human detection. The number of parameters does not correlate to inference speed. For limited use cases, an input image resolution of 64x64 is sufficient. High-level object detecti…

A repository for storing models that have been inter-converted between various frameworks. Supported frameworks are TensorFlow, PyTorch, ONNX, OpenVINO, TFJS, TFTRT, TensorFlowLite (Float32/16/INT8…

Open closed eyes classification. Ultra-fast wink and blink estimation model.

Waving Hand Classification. Ultrafast 1x3x4x32x32 3DConv gesture estimation.

Happy smile classifier. The estimation is done on the entire head, 48x48 pixels, rather than the face.

Phone Usage Classifier (PUC) is a three-class image classification pipeline for understanding how people interact with smartphones.

Ultrafast sitting classification. 32x24 pixels is sufficient for estimating the state of the whole human body.

Pointing gesture classification. Classify whether the finger is pointing near the center of the camera lens. A resolution of 32x32 is sufficient for human hand gesture classification.

Visual only speech detection by lip movement. There are countless situations where you can't hear the audio, and it's really frustrating.

reflection-removal: A model for removing reflections using a single image. The goal is to significantly reduce the amount of calculation required for GAN and remove reflections in real time. The qu…

Command-line utility for splitting MP4 files into multiple clips based on a CSV segment list.

Gaze-LLE-DINOv3: Gaze Target Estimation via Large-Scale Learned Encoders with DINOv3.

HISDF (Human Instance, Skeleton, and Depth Fusion) is a unified model that fuses human instance segmentation, skeletal structure estimation, and depth prediction to achieve holistic human perceptio…

A very simple tool to rewrite parameters such as attributes and constants for OPs in ONNX models. Simple Attribute and Constant Modifier for ONNX.

Tools for simple inference testing using CPU, WebGL, WebGPU. Simple Inference Test for onnxruntime-web.

Simple tool to combine(merge) onnx models. Simple Network Combine Tool for ONNX.

ROI-based Instance Segmentation for Human Detection (CNN)

Simple model Output OP Additional tools for ONNX.