InterActive: Inter-Layer Activeness Propagation

By Lingxi Xie, Liang Zheng, Jingdong Wang, Alan L. Yuille, Qi Tian. 2016

50 citations recorded by the source corpus.

An increasing number of computer vision tasks can be tackled with deep features, which are the intermediate outputs of a pre-trained Convolutional Neural Network. Despite the astonishing performance, deep features extracted from low-level neurons are still below satisfaction, arguably because they cannot access the spatial context contained in the higher layers. In this paper, we present InterActive, a novel algorithm which computes the activeness of neurons and network connections. Activeness is propagated through a neural network in a top-down manner, carrying highlevel context and improving the descriptive power of lowlevel and mid-level neurons. Visualization indicates that neuron activeness can be interpreted as spatial-weighted neuron responses. We achieve state-of-the-art classification performance on a wide range of image datasets.

Subfields: Artificial Intelligence, Biophysics, Computer Vision and Pattern Recognition

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

1. Alan L. Yuille
2. Yi Yang
3. Jiashi Feng
4. Qi Tian
5. Lingxi Xie
6. Shuicheng Yan
7. Jianfei Cai
8. Trevor Darrell

In Biophysics

1. Shuicheng Yan
2. Yuxin Peng
3. Liang Zheng
4. Qi Tian
5. Alan L. Yuille
6. Xiangteng He
7. Jingdong Wang
8. Jianfei Cai

In Computer Vision and Pattern Recognition

1. Yi Yang
2. Alan L. Yuille
3. Luc Van Gool
4. Dacheng Tao
5. Qi Tian
6. Ming-Hsuan Yang
7. Jiashi Feng
8. Xiaogang Wang

Citations in the corpus

• Recent advances in convolutional neural networks (2015, background): “Specifically, stochastic pooling first computes the probabilities p for each region R j by normalizing the activations within the region, i.e ., p i = a i / k j ( a k ) .”
• Genetic CNN (2017, background): “Features extracted from pre-trained neural networks can be generalized to other recognition tasks [36][40].”
• FaceNet2ExpNet: Regularizing a Deep Face Recognition Net for Expression Recognition (2016, background): “In [26], Xie et al. observed that the high-level neurons are exponentially decayed.”
• A survey on deep learning-based fine-grained object classification and semantic segmentation (2017, background): “Some other general deep convolutional feature extractors for image classification include CNN features off-the-shelf [11] , ONE [12] and InterActive [13] .”
• Object-Part Attention Model for Fine-Grained Image Classification (2017, methodology): “CNN Features Our OPAM Approach 97.10 VGGNet InterActive [30] 96.40 VGGNet PBC [57] 96.10 GoogleNet TL Atten [15] 95.76 VGGNet NAC [20] 95.34 VGGNet RIIR [58] 94.01 VGGNet Deep Optimized [51] 91.30 AlexNet SDR [51] 90.50 AlexNet MML [52] 89.45 CNN CNN Feature [59] 86.80 CNN Generalized Max Pooling…”
• SORT: Second-Order Response Transform for Visual Recognition (2017, background): “The intermediate responses of CNNs can be used as effective image descriptors [6], and widely applied to various types of vision applications, including image classification [24][56] and instance retrieval [42][54].”
• SIFT Meets CNN: A Decade Survey of Instance Retrieval (2016, background): “[141] improve the MAC representation [10] by propagating the high-level semantics and spatial context to low-level neurons for improving the descriptive ability of these bottomlayer activations.”
• DisturbLabel: Regularizing CNN on the Loss Layer (2016, background): “CNN also provides efficient visual features for other tasks [4][26][39][40][41].”
• Geometric Neural Phrase Pooling: Modeling the Spatial Co-occurrence of Neurons (2016, background): “They can be used for various types of vision tasks, including image classification [4,37], image retrieval [6,7] and object detection [8].”