MIRACL-VC1

MIRACL-VC1是包括深度和彩色图像的唇读数据集。它可用于多种研究领域，例如视觉识别，人脸检测和生物识别。15个扬声器 (五个男人和十个女人) 位于MS Kinect传感器的平截头中，说出十个单词和十个短语的十倍 (请参见下表)。数据集的每个实例都由颜色和深度图像的同步序列组成 (均为640x480像素)。MIRACL-VC1数据集包含总共3000个实例。MIRACL-VC1 is a lip-reading dataset that includes both depth and color images. It can be applied to a variety of research fields, such as visual recognition, face detection, and biometrics. Fifteen speakers (five males and ten females) were positioned within the flat truncated cone of the MS Kinect sensor, and they uttered ten words and ten phrases, with each being repeated ten times (please refer to the table below). Each instance in the dataset consists of synchronized sequences of color and depth images, both with a resolution of 640×480 pixels. The MIRACL-VC1 dataset contains a total of 3000 instances.

搜集汇总

背景与挑战

背景概述

MIRACL-VC1是一个唇读数据集，包含深度和彩色图像序列，适用于视觉识别和人脸检测等领域。该数据集由15位说话者录制10个单词和10个短语，总计3000个实例，由斯法克斯大学和Valeo于2014年发布。