-
Performance Analysis of a Foreground Segmentation Neural Network Model
Authors:
Joel Tomás Morais,
António Ramires Fernandes,
André Leite Ferreira,
Bruno Faria
Abstract:
In recent years the interest in segmentation has been growing, being used in a wide range of applications such as fraud detection, anomaly detection in public health and intrusion detection. We present an ablation study of FgSegNet_v2, analysing its three stages: (i) Encoder, (ii) Feature Pooling Module and (iii) Decoder. The result of this study is a proposal of a variation of the aforementioned…
▽ More
In recent years the interest in segmentation has been growing, being used in a wide range of applications such as fraud detection, anomaly detection in public health and intrusion detection. We present an ablation study of FgSegNet_v2, analysing its three stages: (i) Encoder, (ii) Feature Pooling Module and (iii) Decoder. The result of this study is a proposal of a variation of the aforementioned method that surpasses state of the art results. Three datasets are used for testing: CDNet2014, SBI2015 and CityScapes. In CDNet2014 we got an overall improvement compared to the state of the art, mainly in the LowFrameRate subset. The presented approach is promising as it produces comparable results with the state of the art (SBI2015 and Cityscapes datasets) in very different conditions, such as different lighting conditions.
△ Less
Submitted 25 May, 2021;
originally announced May 2021.
-
An Hybrid Method for the Estimation of the Breast Mechanical Parameters
Authors:
Diogo Lopes,
António Ramires Fernandes,
Stéphane Clain
Abstract:
There are several numerical models that describe real phenomena being used to solve complex problems. For example, an accurate numerical breast model can provide assistance to surgeons with visual information of the breast as a result of a surgery simulation. The process of finding the model parameters requires numeric inputs, either based in medical imaging techniques, or other measures. Inputs c…
▽ More
There are several numerical models that describe real phenomena being used to solve complex problems. For example, an accurate numerical breast model can provide assistance to surgeons with visual information of the breast as a result of a surgery simulation. The process of finding the model parameters requires numeric inputs, either based in medical imaging techniques, or other measures. Inputs can be processed by iterative methods (inverse elasticity solvers). Such solvers are highly robust and provide solutions within the required degree of accuracy. However, their computational complexity is costly. On the other hand, machine learning based approaches provide outputs in real-time. Although high accuracy rates can be achieved, these methods are not exempt from producing solutions outside the required degree of accuracy. In the context of real life situations, a non accurate solution might present complications to the patient.
We present an hybrid parameter estimation method to take advantage of the positive features of each of the aforementioned approaches. Our method preserves both the real-time performance of deep-learning methods, and the reliability of inverse elasticity solvers. The underlying reasoning behind our proposal is the fact that deep-learning methods, such as neural networks, can provide accurate results in the majority of cases and they just need a fail-safe system to ensure its reliability. Hence, we propose using a Multilayer Neural Networks (MNN) to get an estimation which is in turn validated by a iterative solver. In case the MNN provides an estimation not within the required accuracy range, the solver refines the estimation until the required accuracy is achieved. Based on our results we can conclude that the presented hybrid method is able to complement the computational performance of MNNs with the robustness of iterative solver approaches.
△ Less
Submitted 9 March, 2020;
originally announced March 2020.
-
Parameters identification method for breast biomechanical numerical model
Authors:
Diogo Lopes,
Stéphane Clain,
António Ramires Fernandes
Abstract:
Bio-mechanical breast simulations are based on a gravity free geometry as a reference domain and a nonlinear mechanical model parameterised by physical coefficients. As opposed to complex models proposed in the literature based on medical imagery, we propose a simple but yet realistic model that uses a basic set of measurements easy to realise in the context of routinely operations. Both the mecha…
▽ More
Bio-mechanical breast simulations are based on a gravity free geometry as a reference domain and a nonlinear mechanical model parameterised by physical coefficients. As opposed to complex models proposed in the literature based on medical imagery, we propose a simple but yet realistic model that uses a basic set of measurements easy to realise in the context of routinely operations. Both the mechanical system and the geometry are controlled with parameters we shall identify in an optimisation procedure. We give a detailed presentation of the model together with the optimisation method and the associated discretisation. Sensitivity analysis is then carried out to evaluate the robustness of the method.
△ Less
Submitted 15 July, 2019; v1 submitted 6 December, 2018;
originally announced January 2019.
