TernaryNet: faster deep model inference without GPUs for medical 3D segmentation using sparse and binary convolutions

Heinrich, Mattias P.; Blendowski, Max; Oktay, Ozan

doi:10.1007/s11548-018-1797-4

Cited by 32 publications

(26 citation statements)

References 27 publications

(35 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Overall, the algorithm achieves good performance with 2 to 3% increase in dice score as compared to the existing methods. A related research on pancreas segmentation had been conducted previously using dense connection by Gibson et al [150] and sparse convolutions by Heinrich et al [151], [152], [153]. For multi-organ segmentation (i.e lung, heart, liver, bone) in unlabeled X-ray images, Zhang et al [154] proposed a Task Driven Generative Adversarial Network (TD-GAN) automated technique.…”

Section: Abdomenmentioning

confidence: 99%

Going Deep in Medical Image Analysis: Concepts, Methods, Challenges, and Future Directions

et al. 2019

View full text Add to dashboard Cite

Medical Image Analysis is currently experiencing a paradigm shift due to Deep Learning. This technology has recently attracted so much interest of the Medical Imaging community that it led to a specialized conference in 'Medical Imaging with Deep Learning' in the year 2018. This article surveys the recent developments in this direction, and provides a critical review of the related major aspects. We organize the reviewed literature according to the underlying Pattern Recognition tasks, and further sub-categorize it following a taxonomy based on human anatomy. This article does not assume prior knowledge of Deep Learning and makes a significant contribution in explaining the core Deep Learning concepts to the non-experts in the Medical community. Unique to this study is the Computer Vision/Machine Learning perspective taken on the advances of Deep Learning in Medical Imaging. This enables us to single out 'lack of appropriately annotated large-scale datasets' as the core challenge (among other challenges) in this research direction. We draw on the insights from the sister research fields of Computer Vision, Pattern Recognition and Machine Learning etc.; where the techniques of dealing with such challenges have already matured, to provide promising directions for the Medical Imaging community to fully harness Deep Learning in the future.

show abstract

Section: Abdomenmentioning

confidence: 99%

Going Deep in Medical Image Analysis: Concepts, Methods, Challenges, and Future Directions

et al. 2019

View full text Add to dashboard Cite

show abstract

“…the entire network [10]. However, 3DQ computes one ∆ per layer l, to maintain the variability in weight range values within each layer and avoid weight sparsity [11].…”

Section: Weight Quantizationmentioning

confidence: 99%

“…Even though it is generally accepted that XNOR-Net revolutionized this field [9], there is a significant trade-off between performance and speed, not ideal for medical applications. TernaryNet [10] was the first attempt in medical imaging to create compact and efficient F-CNNs utilizing ternary weights, where a 2D U-Net was employed for the task of per-slice pancreas CT segmentation. However, extending quantization to 3D F-CNNs has yet to be explored.…”

Section: Introductionmentioning

confidence: 99%

3DQ: Compact Quantized Neural Networks for Volumetric Whole Brain Segmentation

Paschali¹,

Gasperini²,

Roy

et al. 2019

Lecture Notes in Computer Science

View full text Add to dashboard Cite

Model architectures have been dramatically increasing in size, improving performance at the cost of resource requirements. In this paper we propose 3DQ, a ternary quantization method, applied for the first time to 3D Fully Convolutional Neural Networks (F-CNNs), enabling 16x model compression while maintaining performance on par with full precision models. We extensively evaluate 3DQ on two datasets for the challenging task of whole brain segmentation. Additionally, we showcase our method's ability to generalize on two common 3D architectures, namely 3D U-Net and V-Net. Outperforming a variety of baselines, the proposed method is capable of compressing large 3D models to a few MBytes, alleviating the storage needs in space-critical applications.

show abstract

“…The residual block diagram is as follows: Figure 3. Residual block Now the residual block is introduced into the feature extraction part of U-Net, so that the feature extraction part adopts the deep network to extract features as abstractly as possible, which is conducive to the subsequent segmentation task [15].…”

Section: U-net Network Based On Residual Blockmentioning

confidence: 99%

“…In the lane detection mentioned in this paper, besides pre-processing the dataset to a certain extent, two u-net [1] networks based on RESNET residual module [2] are redesigned with previous experience based on the simplified deep lab v3p model provided by paddle. Finally, the average results of three networks are adopted to improve the training accuracy.…”

Section: Introductionmentioning

confidence: 99%

Study on Optimized Lane Detection Algorithm based on U-Net

Yao¹,

Zhao²,

Feng³

et al. 2019

Proceedings of the 3rd International Conference on Computer Engineering, Information Science &Amp; Application Technology (ICCI

View full text Add to dashboard Cite

Lane detection has always been one of the important researches in semantic segmentation, but there are many problems in traditional lane detection algorithms, such as the much larger image pixels, the poor detection effect and so on. Based on the U-Net semantics segmentation network model, this paper redesigns two U-Net optimization network models based on RESNET residual module, and puts forward a series of image preprocessing methods aiming at the dataset's much larger pixels and some other problems. In the training process, the training data are adjusted Besides, date cleaning, data enhancement, data exposure and other operations are added. The final training model performs well on Apollos capes dataset.

show abstract

TernaryNet: faster deep model inference without GPUs for medical 3D segmentation using sparse and binary convolutions

Abstract: We present a key enabling technique for highly efficient DCNN inference without GPUs that will help to bring the advances of deep learning to practical clinical applications. It has also great promise for improving accuracies in large-scale medical data retrieval.

Cited by 32 publications

References 27 publications

Going Deep in Medical Image Analysis: Concepts, Methods, Challenges, and Future Directions

Going Deep in Medical Image Analysis: Concepts, Methods, Challenges, and Future Directions

3DQ: Compact Quantized Neural Networks for Volumetric Whole Brain Segmentation

Study on Optimized Lane Detection Algorithm based on U-Net

Contact Info

Product

Resources

About