Improving Computer-Aided Detection UsingConvolutional Neural Networks and Random View Aggregation

“Potentially, larger and deeper convolutional neural networks could be applied to further improve classification performance [59], [60].”

“In order to increase the variation of the training data and to avoid overfitting analogous to the data augmentation approach in [14], [17] and [18], multiple 2D or 2.”

“Potentially, larger and deeper convolutional neural networks could be applied to further improve classification performance [59], [60].”

“This type of cascade classification is similar in spirit to our presented second tier deep classifier (ConvNet), but uses state-of-the-art shallow classifiers (libSVM [47] and sparse linear fusion via the Relevance Vector Machine [48]).”

“Potentially allowing the ConvNet optimization to start from an initialization that is better than starting from Gaussian random parameters [57], [58].”

“Balancing between positive and negative training populations is generally beneficial for training ConvNets when optimizing with logistic regression cost [15], [14].”

“Likewise, the compositional two-stream 2D ConvNet models run on separate spatial (RGB) and temporal (i.e., optical flow field) video frames and achieve the mean accuracy of 87.9% in action classification task, based on a middle scale dataset UCF-101 [50].”

“It is a variation of the previously suggested “DropOut” method [29], [30].”

“This result significantly outperforms the direct 3D “spatial-temporal” ConvNet method [51] at 65.”

“Potentially allowing the ConvNet optimization to start from an initialization that is better than starting from Gaussian random parameters [57], [58].”

“In order to increase the variation of the training data and to avoid overfitting analogous to the data augmentation approach in [14], [17] and [18], multiple 2D or 2.”

“[14], [31] which efficiently trains the ConvNet by using GPU acceleration with the DropConnect modification by [28].”

“In CT colonography (CTC), patients are typically scanned in the prone and supine positions [56], so we obtain two CT volumes per patient study.”

“Additionally, recent work has shown how the implementation of ConvNets can substantially improve the performance of state-of-the-art CADe systems [19]–[22].”

“This strategy has been adapted to computer-aided pulmonary embolism detection ([52]), lung nodule classification ([11], [53]) in CT images and polyp detection in colonoscopy videos ([54], [55]).”

“At present, examples of CADe used in clinical practice include polyp detection for colon cancer screening [9], [10], lung nodule detection for lung cancer screening [11], [12] or breast cancer screening with mammography [13].”

“For instance, [19] proposes an MRI-based knee cartilage segmentation using a triplanar ConvNet.”

“Extensions of ConvNets to 3D have been proposed, but computational cost and memory consumption can be still too high to efficiently implement them on current computer graphics hardware units [23].”

“([20]), effectively circumvents the “curse-of-dimensionality” or “over-fitting” issue in the data-demanding ConvNet training procedures.”

“There were 79 training and 173 testing polyps ; and 22 training and 37 testing polyps ( , considered as large polyps) in our CTC dataset [27].”

“This strategy has been adapted to computer-aided pulmonary embolism detection ([52]), lung nodule classification ([11], [53]) in CT images and polyp detection in colonoscopy videos ([54], [55]).”

“tions of each ROI—an approach similar to [32].”

“At present, examples of CADe used in clinical practice include polyp detection for colon cancer screening [9], [10], lung nodule detection for lung cancer screening [11], [12] or breast cancer screening with mammography [13].”

“These automatically learned filters need no tuning by hand, a major advantage over more traditional CADe approaches [19].”

“At present, examples of CADe used in clinical practice include polyp detection for colon cancer screening [9], [10], lung nodule detection for lung cancer screening [11], [12] or breast cancer screening with mammography [13].”

“This type of cascade classification is similar in spirit to our presented second tier deep classifier (ConvNet), but uses state-of-the-art shallow classifiers (libSVM [47] and sparse linear fusion via the Relevance Vector Machine [48]).”

“This strategy has been adapted to computer-aided pulmonary embolism detection ([52]), lung nodule classification ([11], [53]) in CT images and polyp detection in colonoscopy videos ([54], [55]).”

“Furthermore, detection can help to assess the staging of a patient's disease, and thus has the potential to alter a patient's required treatment regimen [2].”

“We use a soft-tissue window level of [-100, 200 HU] as in [44].”

“At present, examples of CADe used in clinical practice include polyp detection for colon cancer screening [9], [10], lung nodule detection for lung cancer screening [11], [12] or breast cancer screening with mammography [13].”

“Note that this shallow linear cascade approach via new image features, such as Histogram of Oriented Gradients, already significantly surpasses previous state-of-the-art methods [45], [46], [49].”

“, intensity statistics, histogram of oriented gradients (HoG) [6], scale-invariant feature transform (SIFT) [7], Hessian based shape descriptors (such as blobness) [8], etc.”

“This strategy has been adapted to computer-aided pulmonary embolism detection ([52]), lung nodule classification ([11], [53]) in CT images and polyp detection in colonoscopy videos ([54], [55]).”

“In this study, we use three existing CADe systems that have previously been described in the literature: Detection of Sclerotic Spine Metastases: We use a recent CADe method for detecting sclerotic metastases candidates from CT volumes [4], [33] (see Section III-D).”

“ConvNets have a better chance of capturing the “essence” of the imaging data set used for training than do hand-crafted features [16], [6]–[8].”

“Therefore CADe has the potential to greatly reduce the radiologists' clinical workload and to serve as a first or second reader for improved assessment of the disease [3]–[5].”

“The colonic polyp CADe dataset scales are similar: 770 tagged-prep CT scans from multiple medical sites (358 training and 412 validation) in [37], [40]; 180 patients (360 CTC volumes) for training and 202 patients (404 volumes) for testing [38].”

“Note that this shallow linear cascade approach via new image features, such as Histogram of Oriented Gradients, already significantly surpasses previous state-of-the-art methods [45], [46], [49].”

“Potentially, the performance could be improved by adopting a local colonic surface alignment, such as [39], to further guide and constrain our random view sampling procedure.”

“Last, a committee of SVMs [34] is trained on these features.”

“There are a total of 1956 true-positive detections from [26], [25].”

“The colonic polyp CADe dataset scales are similar: 770 tagged-prep CT scans from multiple medical sites (358 training and 412 validation) in [37], [40]; 180 patients (360 CTC volumes) for training and 202 patients (404 volumes) for testing [38].”

“There are a total of 1956 true-positive detections from [26], [25].”

“Axial 2D cross sections of the vertebrae are then divided into sub-segments using a watershed algorithm based on local density differences [33].”

“The colonic polyp CADe dataset scales are similar: 770 tagged-prep CT scans from multiple medical sites (358 training and 412 validation) in [37], [40]; 180 patients (360 CTC volumes) for training and 202 patients (404 volumes) for testing [38].”

“Therefore CADe has the potential to greatly reduce the radiologists' clinical workload and to serve as a first or second reader for improved assessment of the disease [3]–[5].”

“There exist two types of cascaded CADe classification architectures for false positive reduction are two types: 1) extraction of new image features followed by retraining of a classifier on all candidates [39], [38], [6], [20], [40] (from Section II-D) or 2) design of application dependent post-filtering components [4…”

“This is a coarse-to-fine classification approach slightly inspired by other CADe schemes such as presented in [24] although our methods are significantly different.”

“3D flexible tubes [41], ileo-cecal valve [42] and extra-colonic”

“In this study, we use three existing CADe systems that have previously been described in the literature: Detection of Sclerotic Spine Metastases: We use a recent CADe method for detecting sclerotic metastases candidates from CT volumes [4], [33] (see Section III-D).”

“In the abdomen, the most recent work ([46]) shows a 70.”

“3D flexible tubes [41], ileo-cecal valve [42] and extra-colonic”

“It is a variation of the previously suggested “DropOut” method [29], [30].”

“The incorpora- ion of contextual or 3D information is also handled using multi- tream CNNs ( Section 2.4.2 , for example by Barbu et al., 2016 and oth et al., 2016b ).”

“…generating samples via affine transformation (i.e., data augmentation) and then train their network from scratch with the augmented dataset (35, 36, 37, 38); (iii) to use deep models trained over a huge number of natural images in computer vision as ‘off-the-shelf’ feature extractor and then…”

“Implicit features could determine fibroids and polyps [1], and characterize irregularities in tissue morphology such as tumors [2].”