Most convolutional neural networks use pooling layers to gradually ... making R-CNN unsuitable for real-time object detection. And finally, the model is composed of three separate components ...