ContextThe CAT dataset includes over 9,000 cat images. For each image, there are annotations of the head of cat with nine points, two for eyes, one for mouth, and six for ears.ContentThe annotation da
ContextThe aim of this dataset is to provide a simple way to get started with 3D computer vision problems such as 3D shape recognition.Accurate3D point cloudscan (easily and cheaply) be adquired nowda
SVHN is a real-world image dataset for developing machine learning and object recognition algorithms with minimal requirement on data preprocessing and formatting. It can be seen as similar in flavor
AbstractThis data was sourced from the J-EDI dataset of marine debris. The videos that comprise that dataset vary greatly in quality, depth, objects in scenes, and the cameras used. They contain image
introductionThe repository presents Tensorflow 2.0 (Keras) implementation of real-time detection of PPE (e.g., hard hat, safety vest) compliances of workers. The detailed procedures can be found in th
Dataset SummaryPublic database released in conjunction with SCIA 2011, 24-26 May, 2011More than20 000 imageswith20% labeledContains3488traffic signsSequences from highways and cities recorded from mor
MNIST数据集是一个手写阿拉伯数字图像识别数据集,图片分辨率为 20x20 灰度图图片,包含‘0 - 9’ 十组手写手写阿拉伯数字的图片。其中,训练样本 60000 ,测试样本 10000,数据为图片的像素点值,作者已经对数据集进行了压缩。OverviewThe MNIST database of handwritten digits, has a training set of 60,000