Discovering Internal Representations from Object-CNNs Using Population Encoding
Authors:
Preprint
Abstract
In this paper, we provide a method for understanding the internal representations of Convolutional Neural Networks (CNNs) trained on objects. We hypothesize that the information is distributed across multiple neuronal responses and propose a simple clustering technique to extract this information, which we call \emph{population encoding}. The population encoding technique looks into the entrails of an object-CNN at multiple layers of the network and shows the implicit presence of mid-level object part semantics distributed in the neuronal responses. Our qualitative visualizations show that population encoding can extract mid-level image patches that are visually tighter than the patches that produce high single-filter activations. Moreover, our comprehensive quantitative experiments using the object key point annotations from the PASCAL3D+ dataset corroborate the visualizations by demonstrating the superiority of population encoding over single-filter detectors, in the task of object-part detection. We also perform some preliminary experiments where we uncover the compositional relations between the adjacent layers using the parts detected by population encoding clusters. Finally, based on the insights gained from this work, we point to various new directions which will enable us to have a better understanding of the CNN's internal representations.Keywords
All Related Versions
This publication has 15 references indexed in Scilit:
- The application of two-level attention models in deep convolutional neural network for fine-grained image classification2015
- Deep neural networks are easily fooled: High confidence predictions for unrecognizable images2015
- Mid-level deep pattern mining2015
- Understanding deep image representations by inverting them2015
- Rich Feature Hierarchies for Accurate Object Detection and Semantic Segmentation2014
- Analyzing the Performance of Multilayer Neural Networks for Object Recognition2014
- Blocks That Shout: Distinctive Parts for Scene Classification2013
- Adaptive deconvolutional networks for mid and high level feature learning2011
- Detecting People Using Mutually Consistent Poselet Activations2010
- Single Units and Sensation: A Neuron Doctrine for Perceptual Psychology?Perception, 1972