|
|
 |
| |
|
|
|
Statistical Analysis of Video Databases for Deception Detection Tasks |
|
|
|
PP: 1189-1194 |
|
|
|
doi:10.18576/amis/180602
|
|
|
|
Author(s) |
|
|
|
Kanat Kozhakhmet,
Aikumis Omirali,
|
|
|
|
Abstract |
|
|
| Deception detection plays a crucial role in a multitude of fields, spanning from law enforcement to psychology. As such, there is a pressing need for effective tools that can accurately determine the truth. To address this need, this project utilizes the latest technological advancements, conducting a comprehensive statistical analysis on two separate video databases: ”Real-Life Trial Data” (Dataset A) and the ”Experimental Dataset” (Dataset B). The ultimate goal of this investigation is to assess the effectiveness of these datasets in deception detection tasks. The study is guided by three main objectives: first, a comparative content analysis of the two datasets; second, an evaluation of the performance of different deception detection algorithms; and finally, an examination of inherent challenges. Through meticulous statistical procedures, including the use of accuracy, precision, recall, F1-score, and ANOVA tests, this study reveals subtle differences in the content of the two datasets.
|
|
|
|
|
|
|
|
