Enhancement of clustering techniques by coupling clustering tree and neural network: Application to brain tumour segmentation

Abstract

Currently, no classical clustering algorithm is efficient on its own. The predefined number of clusters required for their operation does not consistently produce satisfactory segmentation results. They exhibit cluster instability, are vulnerable to the local optimum trap, and are sensitive to noise and imaging artefacts. Most contributions designed to overcome these drawbacks incorporate prior knowledge such as cluster label information and statistic measures that demand minimal labelled training data. Although these approaches improve the segmentation accuracy, they tend to diminish the advantages of clustering algorithms over the supervised learning methods. This study proposes a shift from the use of a predefined number of clusters to a clustering tree-based method for performance enhancement of classical clustering algorithms. The proposed method is a three-stage algorithm. It begins with the extraction of low-level features from a clustering tree. Clustering trees are sets of labelled clusters of an image at multiple clustering resolutions. The second stage extracts high-level features by coupling the clustering tree to a single-layer feedforward neural network. The third stage is the classification stage, where the basic model of a neural network extracts the tumour from a high-level feature map.Because neither of the neural networks requires training, the proposed method is both fully unsupervised and fully automated and retains all its advantages over supervised methods. A performance evaluation using FLAIR MRI images of brain tumour patients from the BRATS2015 and BRATS2020 databases demonstrates significant performance enhancement over four classical clustering algorithms and two of the four proposed techniques were comparable to deep learning methods.

Brain tumor segmentation and survival time prediction using graph momentum fully convolutional network with modified Elman spike neural network

“…In 2021, Osadebey et al 30 had proposed a new clustering method through a coupling clustering tree for BTS. This introduced algorithm consists of three phases, initially, extracting the low‐level features.…”

“…From the above‐mentioned state‐of‐art techniques, acquiring labeled data is one of the significant obstacles because of the scarcity of annotated datasets. This poses a challenge in collecting the substantial amount of labeled data required for training 27–30 . Moreover, this research required to address the challenges of class imbalance, and heterogeneity 31–35,38 .…”

“…This poses a challenge in collecting the substantial amount of labeled data required for training. [27][28][29][30] Moreover, this research required to address the challenges of class imbalance, and heterogeneity. [31][32][33][34][35]38 Enhancing the survival prediction task was achieved by taking into account the tumor's location in relation to other brain structures like the ventricles, optic nerve fibers, or significant fiber tracts.…”