动物识别
俄勒冈野生动物,野生动物图像集

5.1G

1284

15

俄勒冈野生动物,野生动物图像集

Deep Learning,Online Communities,Image Data,Computer Vision,Multiclass Classification

Classification

俄勒冈野生动物,野生动物图像集前往PC端下载数据

Description

I had to prepare a presentation for a meetup in Portland, OR area and was looking for a fresh data set. I didn't want to use the well known data sets such as digits-mnist, fashion-mnist, iris flowers, etc. I decided to just create my own data set about something that I love to do. Summer 2019 is coming and Oregon is beautiful to explore and while I'm driving I'm always looking for a deer on the side of the route, or eagles on top of the trees, and go to the local safaris to see bears (don't want to find them in a hike lol) and take some pictures. So well there is when the idea came out. "Let's download a data set of Oregon wildlife and get some fun training a model to classify them".  For this I use a google scraper I found on GitHub. Forked the repository, started a new branch, did some adaptations of that code and downloaded the data set. After the presentation I didn't know what to do with the database. It was just resting on my laptop. I decided to load it here on Kaggle and share some notebooks and hoping this data set can be fun to explore for the community!

Content

Here we find a small data set of 14013 images in the folder oregon_wildlife.zip distributed in 20 classes as follows:

folder: baldeagle images: 748   folder: blackbear images: 718  
folder: cougar images: 680  
folder: elk images: 660  
folder: graywolf images: 730   folder: mountainbeaver images: 577  
folder: bobcat images: 696  
folder: nutria images: 701  
folder: coyote images: 736  
folder: columbianblack-taileddeer images: 735  
folder: seals images: 698  
folder: canadalynx images: 717   folder: ringtail images: 588   folder:  redfox images: 759  
folder: grayfox images: 668   folder: virginiaopossum images: 728  
folder: sea_lions images: 726  
folder: raccoon images: 728  
folder: raven images: 656  
folder: deer images: 764

The second file is a sample of the data with just 5 classes already prepossessed with GapCV library in a h5 file with 3531 images distributed as follows:

key: baldeagle images: 748   key: blackbear images: 718  
key: cougar images: 680  
key: elk images: 660  
key: gray_wolf images: 730


发表评论
0评