I have been currently looking into the guides Matlab provides on training cascade method to detect certain objects. My problem is to count certain objects in an image, specifically products in an image such as Twix, Hersheys, etc and categorize them by product. I will be dealing with multiple photos that may have different orientations of where the products are placed so I need a reliable method of recognizing these products and counting them. Is the train an object cascade method the best way for me to solving this problem??
If there is a better/easier way to going about this problem please provide any information. Also, is there a way to use SURF detection in Matlab to solve this problem?