深度学习与像素级标注

• semantic image segmentataion with deep con-volutional nets and fully connected crfs

1. There are two technical hurdles in the application of DCNNs to image labeling tasks: signal down-sampling, and spatial ‘insensitivity’ (invariance).
2. The main difference between our model and other state-of-the-art models is the combination of pixel-level CRFs and DCNN-based ‘unary terms’.
3. DCNN score maps can reliably predict the presence and rough position of objects in an image but are less well suited for pin-pointing their exact outline.

• Parsing Natural Scenes and Natural Language with Recursive Neural Networks

1. Deep Learning in vision applications can find lower dimensional representations for fixed size input images which are useful for classification

• Deep Human Parsing with Active Template Regression

Farabet et al. [9] trained a multiscale convolutional network from raw pixels to extract dense features for assigning the label to each pixel. However, multiple complex post-processing methods were required for accurate prediction. The recurrent convolutional neural network [25] was proposed to speed up scene parsing and achieved the state-of-theart performance. Girshick et al. [12] also proposed to classify the candidate regions by CNN for semantic segmentation. All of these approaches use the CNNs as local or semi-local classifiers either over superpixels or region hypotheses.

All images in these three datasets contain standing people in frontal/near-frontal view with good visibilities of all body parts.