Enhancing Facial Expression Detection in Live Streaming Scenes

Published on August 10, 2022

Just like how the scent of pheromones can affect others, facial expressions have the power to convey messages and influence decision-making. In the world of live streaming media marketing, the facial expressions of anchors play a crucial role in captivating audiences. A team of researchers has come up with an ingenious solution to improve the detection of anchors’ facial expressions in this context. They have developed an efficient feature extraction network by modifying the YOLOv5 model. Using a two-step cascade classifier and recycler, they filtered out invalid video frames and created a dataset specifically for anchor facial expressions. To enhance accuracy and minimize latency, the researchers fused GhostNet and coordinate attention into the YOLOv5 backbone. The results were impressive, with the modified YOLOv5 outperforming the original model in terms of both speed and accuracy on their self-built dataset. This breakthrough opens up exciting possibilities for enhancing user experience and engagement in live streaming marketing. To delve deeper into this research, check out the full article!

Facial expressions, whether simple or complex, convey pheromones that can affect others. Plentiful sensory input delivered by marketing anchors’ facial expressions to audiences can stimulate consumers’ identification and influence decision-making, especially in live streaming media marketing. This paper proposes an efficient feature extraction network based on the YOLOv5 model for detecting anchors’ facial expressions. First, a two-step cascade classifier and recycler is established to filter invalid video frames to generate a facial expression dataset of anchors. Second, GhostNet and coordinate attention are fused in YOLOv5 to eliminate latency and improve accuracy. YOLOv5 modified with the proposed efficient feature extraction structure outperforms the original YOLOv5 on our self-built dataset in both speed and accuracy.

Read Full Article (External Site)

Leave a Reply

Your email address will not be published. Required fields are marked *

You may use these HTML tags and attributes:

<a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <s> <strike> <strong>