roysti10 / vlfeat_py

A python wrapper for the functions in vlfeat. Its not a wrapper on the Vlfeat C API but rather a wrapper on the functions listed in the Vlfeat C API

Vlfeat_py

A python wrapper for the functions in vlfeat. Its not a wrapper on the Vlfeat C API but rather a wrapper on the functions listed in the Vlfeat C API

Features

• Extracts out SIFT Keypoints and Descriptors for an image
• Extracts out Dense SIFT Keypoints and Descriptors for an image
• Extracts the Region seeds and the frames for an MSER filter in an image

Quick Start

Installation

To install the package, run the following command

pip install -e git+https://github.com/lucasace/vlfeat_py.git#egg=vlfeat_py

Usage

To apply any of the supported descriptors for example SIFT, you can use the following code

import cv2
import vlfeat_py as vl

img = cv2.imread("path/to/image.jpg", cv2.IMREAD_GRAYSCALE)
[sift_keypoints, sift_descriptors] = vl.sift_desc(img)

Supported Descriptors

SIFT

[sift_keypoints, sift_descriptors] = vl.sift_desc(img)

Dense SIFT

[dense_sift_keypoints, dense_sift_descriptors] = vl.dsift_desc(img, np.array([0, 0, img.shape[0], img.shape[1]]))

MSER

[region_seeds, frames] = vl.mser_desc(img)

HOG

hog_desc = vl.hog_desc(img, 32)

LBP

lbp_desc = vl.lbp_desc(img, 32)

Contributing

The project is still in its very initial phase, so contributions are always welcome. Refer to TODO.md to maybe pick out something to work on or propose your own feature by opening an issue.

If there is any issue with the package, please open an issue.

License

This project is licensed under the MIT License - see the LICENSE file for details