A package for image similarity search
Project description
DeepSearchLite
DeepSearchLite is a lightweight and versatile package for finding similar items in a large collection of data, adopted from DeepImageSearch. It supports a variety of data types such as images, text, or any other type of data that can be represented by feature vectors. All you need is a suitable encoder to convert the data into a feature vector, and DeepSearchLite will take care of the rest, enabling you to perform similarity searches efficiently and effectively.
Features
- Supports various data types (images, text, etc.) as long as they can be represented by feature vectors
- Easy to use with custom feature encoders
- Fast similarity search using FAISS indexing
- Extensible with support for custom feature extraction and dimensionality reduction functions
- Add new items to the index without the need for re-indexing the entire dataset
Installation
Install DeepSearchLite using pip:
pip install DeepSearchLite
Quickstart
To use DeepSearchLite, you need to provide a custom feature encoder that can convert your data into feature vectors. For example, let's say you have a collection of images and you want to find similar images. You can use a pre-trained deep learning model as a feature encoder. Please refer to the Example folder for a complete example on how to set up DeepSearchLite for image similarity search. For other types of data, you need to provide a suitable encoder that can convert the data into feature vectors.
Customization
DeepSearchLite allows you to customize feature extraction and dimensionality reduction functions. You can provide your custom functions when initializing the SearchSetup
instance:
search_setup = SearchSetup(
image_list=image_list,
feature_extractor=custom_feature_extractor,
dim_reduction=custom_dim_reduction_function
)
Contributing
Contributions to DeepSearchLite are welcome! If you have an idea, bug report, or feature request, please open an issue on the GitHub repository. If you'd like to contribute code, please fork the repository and submit a pull request.
License
DeepSearchLite is released under the MIT License.
Project details
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Source Distribution
Built Distribution
Hashes for DeepSearchLite-0.1.1-py3-none-any.whl
Algorithm | Hash digest | |
---|---|---|
SHA256 | 9c560d47ad5c80718a45604a4fa4f261eedc3a235b0005140430efa283c07e0d |
|
MD5 | 84c192a3aec244edaad75360b6f9c8e0 |
|
BLAKE2b-256 | 5d1c1764385616798972e9540cf8aa17604b4ab2cd46a4208e6c4ccf1333887e |