I want to analyze a playlist of youtube videos for several things: First, I want to automatically analyze the images in the video using a Resnet to detect the objects in the video. Second, at the same time, I want to extract information about the subcriber count, number of links, comments and date of comments for analysis.
does anyone have any ideas or links to give me to help me accomplish this? Do I need to download the videos before I analyze them or is there a way I can play the videos on the webpage and just run an object detector on it.
You can use youtube-dl to download all videos in a playlist.You will have to download the videos first. The option of screencapturing and then running the classifier on the fly is possible but I guess it will we very inefficient.
How are you planning to annotate the data set? Ex - A video could be abut cars or cats or anything. To train a model, you will need to provide a annotated dataset.
I plan to use the files from the dataset Youtube-8M. there are codes to run sementation or object detection on video frames. Iām exploring its viability now.