opencv - Defining a threshold for feature matching in geometrical re-ranking -


i'm implementing cache virtual reality applications: given input image query, return result associated visually similar cached image (so processed query) if distance between query representation , cached image representation lower threshold. our cache relatively small , contains 10k images representations.

we use vlad codes [1] image representation since compact , incredibly fast compute (around 1 ms).

however, has been shown in [2] the distance between query code , images in dataset (the cache in case) different query query, it's not trivial find absolute threshold. in same work it's proposed method object detection applications, not relevant in context (we return similar image, not , images containing query subject).

[3] offers precise method, @ same time it's expensive , returns short lists. it's based on spatial feature matching re-ranking, , if want know more details quoted section @ end of question. i'm not expert in computer vision, step sounds me lot using feature matcher on short-list of top-k elements according image representation , re-rank them based on number of features matched. my first question is: correct?

in our case approach not problem, since of times top-10 similar vlad codes contains query subject, , should spatial matching step on 10 images.

however, @ point have second question: if had problem of deciding absolute threshold image representations (as vlad codes), problem still persists approach? in first case, threshold "the l2 distance between query vlad code , closest vlad code", here instead threshold value represent "the number of features matched between query image , image closest image using vlad codes".

of course second question makes sense if first question positive.

the approach of [3]:

geometrical re-ranking verifies global geometrical consistency between matches (lowe 2004; philbin et al. 2007) short-list of database images returned image search system. here implement approach of lowe (2004) , apply short-list of 200 images. first obtain set of matches, i.e., each descriptor of query image matched 10 closest ones in short-list images. estimate affine 2d transforma- tion in 2 steps. first, hough scheme estimates trans- formation 4 degrees of freedom. each pair of matching regions generates set of parameters “vote” in 4d histogram. in second step, sets of matches largest bins used estimate finer 2d affine transform. images geometrical estimation succeeds returned in first positions , ranked score based on number of inliers. images estima- tion failed appended geometrically matched ones, order unchanged.


Comments

Post a Comment

Popular posts from this blog

c# - Update a combobox from a presenter (MVP) -

i using mvp in project, new in mvp. have 2 comboboxes. when select option in combobox, combobox should filled new data. this action in presenter. view 'view1' in presenter, , introduced combobox1 , combobox2 properties in 'view1', because need 'datasource', 'displaymember', 'valuemember' , 'refresh()' in method below. but, when using pattern, enough send property like public string combobox2 { { return combobox1.selectedvalue.tosstring(); } } into presenter not whole combobox. how can solve problem? public void onselectedindexchangedcombobox1() { if (view1.combobox1.selectedindex == -1) { return; } datatable dt = tools.getdatatable("a path"); var query = (from o in dt.asenumerable() o.field<string>("afield") == farmerview.combobox1.selectedvalue.tostring() orderby o.field<string>("anotherfield") select...
Read more

How to understand 2 main() functions after using uftrace to profile the C++ program? -

i trying uftrace profile following simple c++ program: #include <iostream> class { public: a() {std::cout << "a created" << std::endl;} ~a() {std::cout << "a destroyed" << std::endl;} }; int main() { a; return 0; } the profile result this: # uftrace a.out created destroyed # duration tid function 2.026 [ 4828] | __cxa_atexit(); [ 4828] | main() { [ 4828] | __static_initialization_and_destruction_0() { 89.397 [ 4828] | std::ios_base::init::init(); 0.768 [ 4828] | __cxa_atexit(); 93.029 [ 4828] | } /* __static_initialization_and_destruction_0 */ 94.425 [ 4828] | } /* main */ [ 4828] | main() { [ 4828] | a::a() { 11.104 [ 4828] | std::operator <<(); 10.825 [ 4828] | std::basic_ostream::operator <<(); 24.514 [ 4828] | } /* a::a */ [ 4828] | a::~a() { 0.978 [ 4828] | ...
Read more

How to put a lock and transaction on table using spring 4 or above using jdbcTemplate and annotations like @Transactional? -

i trying put lock on table while writing on table , if happened in between roll-back . trying convert below code lock table test_g1 read; lock table test_g write; -- begin; start transaction; insert test_g1 values(143); insert test_g values(145); select * test_g1; select * test_g; rollback; select * test_g; unlock tables; how convert above code @transactional spring jdbctemplate code? @transactional(rollbackfor={dataaccessexception.class}) public void test(){ jdbctemplate.execute("insert test1 (id, nam) values (4, 'a')"); throw new dataaccessexception("error") { }; } here trying throw error, insert statement should rollback it's not happening . thanks edit-1 attaching code , doing in jdbcdaoimpl.java , mentioned problem comment above test() . app.java package com.cgiri.javabrains.spring4; import org.springframework.beans.factory.annotation.autowired; import org.springframework.context.applic...
Read more