After a first look of the 90 videos of the previous protocols, we could identified both some recurring images and some recurring visual topics, for example surveillance cameras, facial recognition, monitored crowd, and so on. With this visualization is possible to look into the images at multiple levels: at first, to see the different visual clusters; then, it’s possible to see from which portal and from which video the different images belong; and finally, we wanted to link the images used to the approach given in that part of video: by clicking the buttons at the top, the correspondent images are more visibly shown. To do so, we looked again at the videos and we made some screenshots everytime a different type of images was shown. We collected 796 images from the 90 videos and we clusterized them. What we grasped from this research and it’s also interesting to see is that the same image is used in different videos and platforms and also same images are used for different type of approaches.
Using the same videos of the previous protocol, we did some screenshots and we categorized them; we matched the type of image, the portal and the specific video from which the images come from and the approach used to express an opinion within the theme Social Credit System. Thanks to this analysis it’s clear to figure out how the same kind of image can be used to communicate different messages, according to the speaker’s approach. At the same time this process shows the comparison between the different use of images according to the video portals.
Using the same file of the previous protocol, we renamed each screenshot wisely, putting the number of video and the time in which the image Using the same file of the previous protocol, we renamed each screenshot wisely, putting the number of video and the time in which the image appeared. Then, by checking the time sections made for the previous protocol, we were able to say clearly when the image was used and to express which kind of opinion.