Features to Text: A Comprehensive Survey of Deep Learning on Semantic Segmentation and Image Captioning

Figure 5

Dense captioning illustrating multiple annotations with a single forward pass [142].