0

headcount detection with opencv?

java

asked 2017-11-26 19:35:13 -0600

IL416
1 ●1

Hi All, I am trying to build a solution for the headcount from the queue on image with opencv with using Java. I build opencv from the source on Ubuntu. I tried all the cascades that is shipped with the opencv nothing is working because the object does not facing /look the camera necessarily. Only Hog cascade pedestrian detection is giving about 50% result from the side angle. While i will try to train my own cascade, my question is is the possible with Opencv or please direct to right direction if i am following the wrong approach. Your suggestions will be greatly appreciated.

edit retag flag offensive close merge delete

1 answer

Sort by » oldest newest most voted

0

answered 2017-11-27 03:50:29 -0600

berak
32993 ●7 ●81 ●312

updated 2017-11-27 04:23:04 -0600

if you cannot guarantee, that folks are looking stright into the camera, try the following:

use 2 CascadeClassifiers, one with a frontal face detector, and the other with the profile detector

then, you have to do 3 detections, 1st with the frontal, 2nd with the profile, and a 3rd profile detection with the flipped image.

then use IOU to filter out duplicate detections

"While i will try to train my own cascade"

i have some doubts, that you will be able to get better results. (you'd need a ton of data for this)

edit flag offensive delete link

Comments

Thank you @berak I will follow your suggestions. Will post once i have some positive results. Thank you!

IL416 ( 2017-11-27 14:45:12 -0600 )edit

Links

Question Tools

1 follower

Stats

Asked: 2017-11-26 19:35:13 -0600

Seen: 721 times

Last updated: Nov 27 '17

Related questions

Copyright OpenCV foundation, 2012-2018. Content on this site is licensed under a Creative Commons Attribution Share Alike 3.0 license.

Powered by Askbot version 0.10.2

here is how