Yes. The GRIDSMART Box discriminates bikes, and often other vulnerable road users (VRUs), from motor vehicles with image-based classification using a deep neural network. In the Box, GRIDSMART errs on the side of caution, allowing a modest false positive (FP) rate to ensure no bicyclists are missed. Since the discrimination is image-based, many bicycle-like motor vehicles (e.g., motorcycles) will be detected and sometimes counted as bikes. Note, however, that motorcycles and other FP detections clear the intersection much more quickly than bicyclists, greatly reducing any potential impact on efficiency. The net efficiency improvement over stop-line discrimination methods is significant.