Poqemon-QoE-Dataset
This work is proposed to describe a new dataset that contains many QoE Influence Factors (QoE IFs) and subjective Mean Opinion Scores(MOS). This dataset is collected during the PoQeMoN project (https://www.ip-label.fr/performance-wire/projet-collaboratif-platform-quality-evaluation-mobile-networks/) where a crowd measurement platform is implemented to assess YouTube end user’s QoE in mobile environments (UMTS, HSPA, LTE, etc.). This dataset concerns collecting a lot of QoE IFs using a VLC media player. The used platform consists on an Android-based applica- tion developed and installed on the end user device (phones and tablets). The testbed experiment mainly consists of the following elements:
• A dedicated mobile application has been developed for experimentation.
• The evaluation was performed at different locations (train station, walking, etc.).
• Users were trained to perform the video session tests.
• Nine devices are used by participants to achieve tests. The devices have different characteristics like screen size, android version and Cpu.
• Several types of videos were used (e.g. sport, movie trailer, documentary, news, music,... etc.).
• Four several mobile networks were tested (Orange, SFR, Bouyegues and Free)
The dataset was built from a crowdsourcing campagn test where 1560 samples covering 29 Quality of Experience Impact Factors (QoE IFs). The used videos are of different types/complexities.
The geographical location of the testbed was the LiSSi laboratory around Paris city in France. 62 testers prticipated in the test campaign. All of them were researchers and students from different disciplines aged 20 to 37 years with few or no experience with video assessment experimentation.
You can download the dataset and find more details here:
https://github.com/Lamyne/Poqemon-QoE-Dataset