YOLOX is an anchor-free version of YOLO with strong performance for object detection. In this repository I train Yolox on WiderFace dataset with Google Colab to develop a face detection algorithm. (back to top) Look at Jupyter Notebooks to see training details and how to make this algorithm from scratch. cd yolox-face-detection.

Ultralytics YOLOv8, developed by Ultralytics, is a cutting-edge, state-of-the-art (SOTA) model that builds upon the success of previous YOLO versions and introduces new features and improvements to further boost performance and flexibility.

The YOLOX network incorporating the attention mechanism is proposed for the problem of wrong detection and omission of YOLOX in the detection of small faces at low resolution. The proposed method were evaluated on some open benchmark face datasets, promising results were obtained.

Based on YOLO detector, the YOLOX model adopts anchor-free head and conducts other advanced detection techniques including decoupled head and the leading label assignment strategy SimOTA to achieve state-of-the-art results across a large scale range of models.

Contribute to xialuxi/yolox-face-landmarks development by creating an account on GitHub.

根据官网的介绍，Wider Face数据集总共有32,203图片进行了人脸标注，总共标注了393,703个人脸数据。 对于每张人脸附带有详细的信息，包括box坐标（左上角+wh）blur（模糊程度）, expression（表情）, illumination（光照）, occlusion（遮挡）, pose（姿态）。 在数据集中，根据事件场景的类型分为了61个类。 接着根据每个类别按照 40% / 10% / 50%的比例划分到训练集，验证集以及测试集中，根据数据统计结果，训练集中有12280张图片，验证集有3226张图片。 …

YOLOX is a high-performance anchor-free YOLO, exceeding yolov3~v5 with MegEngine, ONNX, TensorRT, ncnn, and OpenVINO supported. YOLOXDetect-Pip : This repo is a packaged version of the YOLOX for easy installation and use.

Aug 23, 2023 · To tackle this issue, we present a novel network structure that builds upon CE-FPN to enhance the performance of YoloX network in detecting smaller objects wearing mask targets in complex environments such as stations and subways.

We’re on a journey to advance and democratize artificial intelligence through open source and open science.

To satisfy the required recognition accuracy from a far distance, an adjusted attention-based YOLOX for face recognition (AA-YOLOX-FR) is designed by an appropriate segmentation of the original image, so that faces' pixels (e.g., 20 × 20 at 15m) effectively train, validate, and test.