CC-NET: Image Complexity Guided Network Compression for Biomedical Image Segmentation

@article{Mishra2019CCNETIC,
  title={CC-NET: Image Complexity Guided Network Compression for Biomedical Image Segmentation},
  author={Suraj Mishra and Peixian Liang and Adam Czajka and Danny Ziyi Chen and Xiaobo Sharon Hu},
  journal={2019 IEEE 16th International Symposium on Biomedical Imaging (ISBI 2019)},
  year={2019},
  pages={57-60}
}
Convolutional neural networks (CNNs) for biomedical image analysis are often of very large size, resulting in high memory requirement and high latency of operations. Searching for an acceptable compressed representation of the base CNN for a specific imaging application typically involves a series of time-consuming training/validation experiments to achieve a good compromise between network size and accuracy. To address this challenge, we propose CC-Net, a new image complexity-guided CNN… Expand
Image Complexity Guided Network Compression for Biomedical Image Segmentation
  • Suraj Mishra, Danny Z. Chen, X. Sharon Hu
  • Computer Science, Engineering
  • ArXiv
  • 2021
TLDR
The proposed framework utilizes data complexity and network architecture to quickly estimate a compressed model which does not require network training, and mapping the dataset complexity to the target network accuracy degradation caused by compression enables the final accuracy for different network sizes to be predicted. Expand
Obesity May Be Bad: Compressed Convolutional Networks for Biomedical Image Segmentation
Compressed Highly parameterize convolutional neural network (CNN) is essential for practical biomedical image segmentation applications. However, the accuracy of compressed CNN based on currentExpand
OFF-eNET: An Optimally Fused Fully End-to-End Network for Automatic Dense Volumetric 3D Intracranial Blood Vessels Segmentation
TLDR
An Optimally Fused Fully end-to-end Network (OFF-eNET) for automatic segmentation of the volumetric 3D intracranial vascular structures and outperformed the state-of-the-art counterparts. Expand
FCN-Based DenseNet Framework for Automated Detection and Classification of Skin Lesions in Dermoscopy Images
TLDR
A novel FCN-based DenseNet framework that is composed of dense blocks that are merged and connected via the concatenation strategy and transition layer and employs hyper-parameters optimization techniques to reduce network complexity and improve computing efficiency is proposed. Expand
DANCE: DAta-Network Co-optimization for Efficient Segmentation Model Training and Inference
  • Chaojian Li, Wuyang Chen, +4 authors Yingyan Lin
  • Computer Science
  • ArXiv
  • 2021
TLDR
DANCE is proposed, general automated DAta-Network Co-optimization for Efficient segmentation model training and inference that can achieve “all-win” towards efficient segmentation (reduced training cost, less expensive inference, and better mean Intersection-over-Union (mIoU). Expand
Exploring The Limits Of Data Augmentation For Retinal Vessel Segmentation
TLDR
It is shown that the performance of the U-Net model can be increased dramatically and the reliance on heavy data augmentation in order to achieve better performance is relied on. Expand
A Fast and Efficient FPGA-based Level Set Hardware Accelerator for Image Segmentation
Level set algorithm has been widely used for image segmentation due to its high accuracy. In addition, compared to the deep learning-based image segmentation methods, the level set algorithm can beExpand
Remote sensing image recognition based on dual-channel deep learning network
TLDR
This paper proposes a recognition framework based on dual-channel deep learning (DCDL) to mine global feature information and local feature information at the same time and can see that the proposed algorithm has achieved higher recognition accuracy. Expand
Semantic Segmentation of Marine Remote Sensing Based on a Cross Direction Attention Mechanism
TLDR
The proposed cross attention mechanism of exponential operation combined with multi-scale convolution algorithm is significantly improved, and proved the powerful performance of the algorithm in the semantic segmentation of marine remote sensing images. Expand
Open Source Iris Recognition Hardware and Software with Presentation Attack Detection
  • Z. Fang, A. Czajka
  • Computer Science
  • 2020 IEEE International Joint Conference on Biometrics (IJCB)
  • 2020
TLDR
A lightweight image complexity-guided convolutional network for fast and accurate iris segmentation, domain-specific human-inspired Binarized Statistical Image Features (BSIF) to build an iris template, and to combine 2D and 3D features for PAD are proposed. Expand
...
1
2
...

References

SHOWING 1-10 OF 18 REFERENCES
U-Net: Convolutional Networks for Biomedical Image Segmentation
TLDR
It is shown that such a network can be trained end-to-end from very few images and outperforms the prior best method (a sliding-window convolutional network) on the ISBI challenge for segmentation of neuronal structures in electron microscopic stacks. Expand
Suggestive Annotation: A Deep Active Learning Framework for Biomedical Image Segmentation
TLDR
A deep active learning framework that combines fully convolutional network (FCN) and active learning to significantly reduce annotation effort by making judicious suggestions on the most effective annotation areas is presented. Expand
Cascaded Fully Convolutional Networks for automatic prenatal ultrasound image segmentation
TLDR
A cascaded framework for fully automatic US image segmentation that can bridge severe boundary incompleteness and achieves the best segmentation accuracy when compared with state-of-the-art methods is proposed. Expand
Deep Contextual Networks for Neuronal Structure Segmentation
TLDR
A deep contextual network is proposed here by leveraging multi-level contextual information from the deep hierarchical structure to achieve better segmentation performance and can potentially facilitate the automatic connectome analysis from EM images with less human intervention effort. Expand
Deep Compression: Compressing Deep Neural Network with Pruning, Trained Quantization and Huffman Coding
TLDR
This work introduces "deep compression", a three stage pipeline: pruning, trained quantization and Huffman coding, that work together to reduce the storage requirement of neural networks by 35x to 49x without affecting their accuracy. Expand
Pruning Convolutional Neural Networks for Resource Efficient Inference
TLDR
It is shown that pruning can lead to more than 10x theoretical (5x practical) reduction in adapted 3D-convolutional filters with a small drop in accuracy in a recurrent gesture classifier. Expand
Gland segmentation in colon histology images: The glas challenge contest
TLDR
An overview to the Gland Segmentation in Colon Histology Images Challenge Contest (GlaS) held at MICCAI'2015 is provided, along with the method descriptions and evaluation results from the top performing methods. Expand
MobileNets: Efficient Convolutional Neural Networks for Mobile Vision Applications
TLDR
This work introduces two simple global hyper-parameters that efficiently trade off between latency and accuracy and demonstrates the effectiveness of MobileNets across a wide range of applications and use cases including object detection, finegrain classification, face attributes and large scale geo-localization. Expand
MorphNet: Fast & Simple Resource-Constrained Structure Learning of Deep Networks
TLDR
MorphNet iteratively shrinks and expands a network, shrinking via a resource-weighted sparsifying regularizer on activations and expanding via a uniform multiplicative factor on all layers, which is scalable to large networks, adaptable to specific resource constraints, and capable of increasing the network's performance. Expand
Image complexity and spatial information
  • Honghai Yu, Stefan Winkler
  • Computer Science
  • 2013 Fifth International Workshop on Quality of Multimedia Experience (QoMEX)
  • 2013
TLDR
It is found that compression-based complexity of an image normally increases with decreasing resolution, and spatial information (SI) measures strongly correlate with compression- based complexity measures. Expand
...
1
2
...