Published on 30 Jan. 2026
Last edit 2 September 2026 
By Xiaoyuan Gao

一直以来我都很喜欢收集各种各样的图像，在ArtEZ上学时做过一个项目，对比人和机器怎么描述陌生图像；我比较偏心,觉得人的描述更有趣，但机器说的东西也很特别——非常字面,有时候识别错了反而更有意思。去年我又开始鼓捣图像识别和裁切，从一些public domain档案网站下了一堆日常照片，并进行了很主观地剪切分类：鸟啊、杯子啊、人站着的样子、等等。这显然是一个体力活……虽然不知道最后这个尝试会成为什么，我姑且就慢慢做下去好了。以下是一部分截图和分类。

Collecting all types of images has been my side hobby since I was studying in ArtEZ, started a project when for my design research class, to compare how humans and "machines" describe an image they have never seen and have no idea of the original context. I think I must been biased and I think that the description written by human is always more interesting to me. Although, it is also interesting to look at what the machine describes an image, sometimes it only recognize something very superfacial, and the most fun part is when it does recoginize something wrong... Since last year I started to restart this initiatives again, although this time is more focusing certain type images i find on archive websites, images might related daily life, and i manually cut-out the things I recognize—birds, cups, humans figures, etc. Obviously this is some time-consuming work, and I don't know where it is going, see below, some of my cut-outs from the collections:

Below are some cropouts from the YOLO model—it identified and extracted these from the archive photos.