I've developed an action recognition algorithm using depth action sequences. I wanna implement the algorithm in a surveillance camera which captures rgb videos. Would the algorithm be able to recognize actions regarding the data?
Can you please elaborate your question? Are you asking whether the algorithm that you have developed for depth data can be applied directly for surveillance (RGB) videos? If this is your question, then the answer depends on the features that you are extracting from your depth sequences for the recognition task. If there is any feature that is directly or indirectly dependent on depth then naturally you cannot apply the algorithm 'directly' on RGB videos.