I want to do scenario description of road. If there are cars, humans infront of the blind person it should describe the scenario and output should be consumable by blind person.
Question: Should it be done on Video or static image ? If we do on video, can we just take the first few frames and describe it. Will it involve Object detection models, semantic segmentation model ?
Question: I also want to calculate the distance of the object from the person using RGBD images/videos. Any pointers to this use case will really help. Which model would ideally be suitable for outdoors like roads.