Name
Abstract | ||
|---|---|---|
In this work, a system that determines the following-up status of spectators, from video scenes that belonging to a presentation to a group, by evaluating the positions of people in a group and a person in stage, was designed and performed. System is evaluating images from at least two cameras synchronously. From images that taken from one of the cameras, the position of the person in the stage, is being determined with Histogram of Oriented Gradients method, and being tracked with Mean-Shift method. On the images that taken from the other camera, with the Haar based face detection method, the faces of the spectators are being detected and tracked. In this work, to determine the direction of spectators looking directions, a basic and effective method that divides the face of tracked face to two, and compares the rate of skin colored pixels on both sides, has been developed. For every spectator, the looking direction and the position of the person on the stage are being matched and it is determined if the spectators are looking at the target person. |
Year | Venue | Field |
|---|---|---|
2017 | SIU | Computer vision,Histogram,Colored,Computer graphics (images),Effective method,Haar,Computer science,Histogram of oriented gradients,Pixel,Artificial intelligence,Face detection |
DocType | Citations | PageRank |
Conference | 0 | 0.34 |
References | Authors | |
0 | 2 | |
Name | Order | Citations | PageRank |
|---|---|---|---|
| M. Elif Karsligil | 1 | 73 | 13.69 |
| Furkan Kaynar | 2 | 0 | 0.68 |