Abstract

Contour detection has been a fundamental component in many image segmentation and object detection systems. Most previous work utilizes low-level features such as texture or saliency to detect contours and then use them as cues for a higher-level task such as object detection. However, we claim that recognizing objects and predicting contours are two mutually related tasks. Contrary to traditional approaches, we show that we can invert the commonly established pipeline: instead of detecting contours with low-level cues for a higher-level recognition task, we exploit objectrelated features as high-level cues for contour detection. We achieve this goal by means of a multi-scale deep network that consists of fifive convolutional layers and a bifurcated fully-connected sub-network. The section from the input layer to the fififth convolutional layer is fifixed and directly lifted from a pre-trained network optimized over a largescale object classifification task. This section of the network is applied to four different scales of the image input. These four parallel and identical streams are then attached to a bifurcated sub-network consisting of two independentlytrained branches. One branch learns to predict the contour likelihood (with a classifification objective) whereas the other branch is trained to learn the fraction of human labelers agreeing about the contour presence at a given point (with a regression criterion). We show that without any feature engineering our multiscale deep learning approach achieves state-of-the-art results in contour detection