image processing - feature detectors and descriptors comparison -


there several kinds of detectors , descriptors, sift, surf, fast. wonder eligible real-time applications? best or better?

and furthermore, harris-laplacian dectector still useful when have above three? better them?

i can advise use hessian-affine , mser detection, if need invariance different factors (e.g., viewpoint change) or fast, if need real time. fast doing similar job harris, faster.

you can "local invariant feature detectors: survey", , "a comparison of affine region detectors" many detectors tested , described well.

update: "wxbs: wide baseline stereo generalizations" extended benchmark of novel , classical detectors , descriptors.

second, description part slower detection, real-time have use gpu or binary descriptor brief or freak.

update2: "hpatches (homography patches) dataset , benchmark" , corresponding workshop @ eccv 2016. http://www.iis.ee.ic.ac.uk/computervision/descrworkshop/index.html .

update3: "comparative evaluation of hand-crafted , learned local features" descriptors (and bit detectors) evaluation on large-scale 3d reconstruction task cvpr 2017 .


Comments

Popular posts from this blog

java - activate/deactivate sonar maven plugin by profile? -

python - TypeError: can only concatenate tuple (not "float") to tuple -

java - What is the difference between String. and String.this. ? -