I am creating a signboard translation application from scratch. I have images of signboards where there are multiple texts and I have the corresponding set of coordinates of bounding boxes for multiple texts. I want to create a regression model which will try to predict the coordinates if there is some text in the image. I am really stuck at a place. In some cases, I have multiple words in the image, so each word will have its own set of coordinates. So, how can I make a model such that if there is a single word then it will output single set of coordinates, but if there are 5 words then it should give me 5 set of coordinates? The number of output may vary with each image. What kind of neural net should I use? I don't want to use sliding window approach. Please help me out.