Performing feature selection on a small number of instances with high-dimensional datasets poses a needed challenge in preventing over-fitting. To address this issue, this paper proposes a sequential transfer-learning...
详细信息
ISBN:
(数字)9798350308365
ISBN:
(纸本)9798350308372
Performing feature selection on a small number of instances with high-dimensional datasets poses a needed challenge in preventing over-fitting. To address this issue, this paper proposes a sequential transfer-learning approach combined with a multi-objective genetic algorithm (STMO-GA) for feature selection. Firstly, for the multi-objective component of our method, we employ a Non-dominated Sorting Genetic Algorithm (NSGA-II) to generate a Pareto front. Then, features are ranked based on their number of appearances in the same Pareto front. Next, during the sequential knowledge transfer process, the ranked features are iteratively selected until a predetermined
$n$
number of features remains. This feature subspace is further refined by a k-fold cross-validation operation, starting from the rank-one feature, to determine the cut-off of the
$n$
features that will remain. Comparative evaluations against both GA-based as well as traditional feature selection methods demonstrate that the proposed method achieves superior classification accuracy, while retaining the smallest number or a comparable number of features.
Implicit neural representations and neural rendering have gained increasing attention for bathymetry estimation from sidescan sonar (SSS). These methods incorporate multiple observations of the same place from SSS dat...
详细信息
Twin-to-twin transfusion syndrome is a serious prenatal condition whose treatment employs the use of a fetoscope and optical laser for targeted vessel ablation. During surgery, challenges such as a limited field-of-vi...
详细信息
ISBN:
(数字)9798350313338
ISBN:
(纸本)9798350313345
Twin-to-twin transfusion syndrome is a serious prenatal condition whose treatment employs the use of a fetoscope and optical laser for targeted vessel ablation. During surgery, challenges such as a limited field-of-view and inconsistent lighting can be addressed by field-of-view expansion. As foundation, we draw upon a state-of-the-art method that utilizes a U-Net segmentation model to mitigate visible in utero inconsistencies. We adapt this method for real-time performance by implementing a feature-based stitching algorithm. The proposed framework includes a loop closure and map correction procedure to help reduce the accumulated drifting error. During evaluation on simulated fetoscopic movements, the chosen feature extractor, SIFT, permits an average homography estimation time of 0.07s and a failure rate of 4%. Homography parameters such as scaling, translation, and rotation are isolated to identify obstacles caused by certain camera movements. The framework is evaluated using two phantom placentas in different environmental conditions and submerged underwater. Qualitative and quantitative results are presented as well as an analysis of the map correction procedure. These evaluations help recognize limitations of the framework that can be investigated and addressed in future research.
Deep learning has shown promising results for multiple 3D point cloud registration datasets. However, in the underwater domain, most registration of multibeam echo-sounder (MBES) point cloud data are still performed u...
详细信息
ISBN:
(数字)9798350384574
ISBN:
(纸本)9798350384581
Deep learning has shown promising results for multiple 3D point cloud registration datasets. However, in the underwater domain, most registration of multibeam echo-sounder (MBES) point cloud data are still performed using classical methods in the iterative closest point (ICP) family. In this work, we curate and release DotsonEast Dataset, a semi-synthetic MBES registration dataset constructed from an autonomous underwater vehicle in West Antarctica. Using this dataset, we systematically benchmark the performance of 2 classical and 4 learning-based methods. The experimental results show that the learning-based methods work well for coarse alignment, and are better at recovering rough transforms consistently at high overlap (20-50%). In comparison, GICP (a variant of ICP) performs well for fine alignment and is better across all metrics at extremely low overlap (10%). To the best of our knowledge, this is the first work to benchmark both learning-based and classical registration methods on an AUV-based MBES dataset. To facilitate future research, both the code and data are made available online.
1
In response to stringent emission reduction targets imposed by theInternational Maritime Organization (IMO) and the European Green Deal\'s Fit for55 legislation package, the maritime industry has shifted its focus...
详细信息
In response to stringent emission reduction targets imposed by the International Maritime Organization (IMO) and the European Green Deal's Fit for 55 legislation package, the maritime industry has shifted its focu...
详细信息
ISBN:
(数字)9798350362077
ISBN:
(纸本)9798350362084
In response to stringent emission reduction targets imposed by the International Maritime Organization (IMO) and the European Green Deal's Fit for 55 legislation package, the maritime industry has shifted its focus towards decarbonization. While significant attention has been placed on vessels exceeding 5,000 gross tons (GT), emissions from coastal and short sea shipping, amounting to approximately 13% of global shipping transportation and 15% within the European Union (EU), have not received adequate consideration. This abstract introduces the Zero Emission Sea Transporter (ZEST) project, designed to address this issue by developing a zero-emissions multi-purpose catamaran for short sea routes, shown in Figure 1.
Plant-derived crop residue on soil surface provides many important advantages including preventing erosion and conserving soil moisture. In that sense, making accurate determination on the percent of crop residue cove...
详细信息
ISBN:
(纸本)9781728190495
Plant-derived crop residue on soil surface provides many important advantages including preventing erosion and conserving soil moisture. In that sense, making accurate determination on the percent of crop residue cover using RGB images can be a fundamental tool in protecting the soil. In our research, we approach the determination of such percentages as a classification problem, and in this paper, we compare two of these approaches. Both approaches relied on support vector machines (SVM) as the classifier of choice, and the same set of features, which were selected in our previous studies on the same topic. In this paper we developed a SVM ensemble with a hierarchical structure and compared it against a single, multi-class SVM classifier. In the SVM ensemble framework, four two-class SVMs and one five-class SVM were combined in sequence to better separate adjacent levels of residue cover. The rationale of the ensemble was to allow each of the two-class SVMs to find the hyperplanes that maximize the margin between the corresponding two consecutive classes. Then, based on the distance of the samples to these hyperplanes, probabilistic estimates of the data-point belonging to the class were computed and added as extra inputs for the last SVM. In order to enhance the performance of the ensemble, other considerations such as the use of Grid Search method for optimizing the hyperparameters were employed in the tuning of the SVMs. Numerical experiments were conducted over a dataset of 4,400 images, which were collected from 88 locations in 40 row crop fields in five Missouri counties between mid-April and early July in 2018 and 2019. The images were collected using a camera mounted on a tripod, with a spatial resolution of 0.014 cm pixel- 1 GSD (Ground Sampling Distance). The experiments highlighted the better performance of the proposed hierarchical ensemble classifier, which achieved a cross-validation accuracy of 86.3 % vs an accuracy of 80.4 % for the single SVM, whil
The correct interpretation of breast density is important in the assessment of breast cancer risk. AI has been shown capable of accurately predicting breast density, however, due to the differences in imaging characte...
详细信息
Promptable segmentation foundation models have emerged as a transformative approach to addressing the diverse needs in medical images, but most existing models require expensive computing, posing a big barrier to thei...
详细信息
Increasing evidence shows that flaws in machine learning (ML) algorithm validation are an underestimated global problem. Particularly in automatic biomedical image analysis, chosen performance metrics often do not ref...
详细信息
暂无评论