To all my computer vision friends working on real-world applications with messy image data, I just open-sourced a Python library you may find useful!<\/p>\n
CleanVision audits any image dataset to automatically detect common issues such as images that are blurry, under\/over-exposed, oddly sized, or near duplicates of others. It\u2019s just 3 lines of code to discover what issues lurk in your data before you dive into modeling, and CleanVision can be used for any<\/strong> image dataset \u2014 regardless of whether your task is image generation, classification, segmentation, object detection, etc.<\/p>\n from cleanvision.imagelab import Imagelab imagelab = Imagelab(data_path=”path_to_dataset”) imagelab.find_issues() imagelab.report() <\/p>\n As leaders like Andrew Ng and OpenAI have lately repeated: models can only be as good as the data they are trained on. Before diving into modeling, quickly run your images through CleanVision to make sure they are ok \u2014 it\u2019s super easy!<\/p>\n Github: https:\/\/github.com\/cleanlab\/cleanvision<\/a><\/p>\n Disclaimer: I am affiliated with Cleanlab.<\/p>\n<\/div>\n submitted by \/u\/jonas__m <\/a>
[link]<\/a><\/span> [comments]<\/a><\/span><\/p>
