Dense SIFT
Simple demo of dense SIFT feature descriptors extraction.
By Laurent Ittiitti@usc.eduhttp://jevois.orgGPL v3
 Language: C++Supports mappings with USB output: YesSupports mappings with NO USB output: No 
 Video Mapping:   YUYV 288 120 5.0 YUYV 160 120 5.0 JeVois DenseSift
 Video Mapping:   GREY 128 117 5.0 YUYV 160 120 5.0 JeVois DenseSift

Module Documentation

Compute SIFT keypoint descriptors on a regular grid over the input image.

This module is useful when using JeVois as a pre-processor, delivering a dense array of keypoint descriptors to a host computer, where the array is disguised as a grayscale video frame. Upon receiving the array of descriptors, the host computer can further process them. For example, the host computer may compute camera motion in space by matching descriptors across successive frames, or may attempt to detect and identify objects based on the descriptors.

Beware that changing the values for the step and binsize parameters changes the output image size, so you need to adjust your video mappings accordingly. Hence, setting those parameters is best done once and for all in the module's optional params.cfg or script.cfg file.

This module can either have a color YUYV output, which shows the original camera image, keypoint locations, and descriptor values; or a greyscale output, which is just the descriptor values.

This algorithm is implemented using the VLfeat library. It is quite slow, maybe because this library is a bit old and appears to be single-threaded.

ParameterTypeDescriptionDefaultValid Values
(DenseSift) stepunsigned intKeypoint step (pixels)11-
(DenseSift) binsizeunsigned intDescriptor bin size8-
Detailed docs:DenseSift
Copyright:Copyright (C) 2016 by Laurent Itti, iLab and the University of Southern California
License:GPL v3
Distribution:Unrestricted
Restrictions:None
Support URL:http://jevois.org/doc
Other URL:http://iLab.usc.edu
Address:University of Southern California, HNB-07A, 3641 Watt Way, Los Angeles, CA 90089-2520, USA