Course Content

Unit 1: Profiling Image Data

Welcome to the unit focusing on image data features. In this unit, we’ll delve into features that derive from attributes intrinsic to images themselves, such as color, scale, and brightness. These features are relevant for all computer vision datasets, regardless of your dataset’s specific task, as they are independent of label-based attributes like bounding boxes or object labels.


About This Unit

In each of the initial three lessons, we’ll focus on a unique image data feature, exploring its significance and the potential discrepancies that can be revealed through its analysis. We’ll also delve into strategies for rectifying these issues to enhance your model’s performance. Our fourth lesson takes a deeper dive into a practical solution—data shuffling. This technique is particularly beneficial in addressing challenges related to the underepresentation of a certain image type or object class in your training set.

Let’s dive in and explore these key image features.  

Add Your Heading Text Here
					from transformers import AutoFeatureExtractor, AutoModelForImageClassification

extractor = AutoFeatureExtractor.from_pretrained("microsoft/resnet-50")

model = AutoModelForImageClassification.from_pretrained("microsoft/resnet-50")