Image Vectorizers

Image Vectorizers in Unbody transform visual data into numerical vectors, facilitating image search and comparison tasks.

In this documentation, we provide an overview of the available image vectorization modules, their operational models, and use cases. we have also included guidance on selecting the appropriate vectorizer for different types of image data.

img2vec-neural

• Overview: The modal transforms images into vectors using the resnet50 model.
• About the Model: Resnet50 is a widely used convolutional neural network for image classification and feature extraction.
• Third-Party Management: The model operates locally. It is not managed by a third party.
• Strengths: It is efficient and effective for feature extraction from images.
• Limitations: The model may be resource-intensive for large image datasets.
• Best For: It is best for in-depth analysis and feature extraction from images.
• Production Status: In production.
• Available Models: resnet50 (keras and pytorch versions).