An experimental evaluation of echo state network for colour image segmentation

Abstract

Image segmentation refers to the process of dividing an image into multiple regions which represent meaningful areas. Image segmentation is an essential step for most image analysis tasks such as object recognition and tracking, pattern recognition, content-based image retrieval, etc. In recent years, a large number of image segmentation algorithms have been developed, but achieving accurate segmentation still remains a challenging task. Recently, reservoir computing (RC) has drawn much attention in machine learning as a new model of recurrent neural networks (RNN). Echo State Network (ESN) represents one efficient realization of RC, which is initially designed to facilitate learning in Recurrent Neural Networks. In this paper we investigate the viability of ESN as feature extractor for pixel classification based colour image segmentation. Extensive experiments are conducted on real world colour image datasets and the global ESN reservoir parameters are varied to identify their operating ranges that allow the use of the reservoir nodes internal activations as new pixel features for the colour image segmentation task. A simple feed forward neural network is used to realize the ESN readout function and classify these new features. The experimental results show that the proposed method achieves high performance image segmentation comparing with state-of-the-art techniques. In addition, a set of empirically derived guidelines for setting the reservoir global parameters are proposed.

Echo state network‐based feature extraction for efficient color image segmentation

“…The present work extends two previous works by the authors 49,50 . In Reference 49, the study of the effect of ESN parameters on color image segmentation performance was limited to a reduced set of the ESN reservoir parameters only, which are the density of connectivity between reservoir nodes, the spectral radius, and the reservoir size.…”

“…The present work extends two previous works by the authors. 49,50 In Reference 49, the study of the effect of ESN parameters on color image segmentation performance was limited to a reduced set of the ESN reservoir parameters only, which are the density of connectivity between reservoir nodes, the spectral radius, and the reservoir size. However, this new work further extends the study of the influence of other parameters on the quality of the segmentation by including: the effect of the selected color space, the order in which inputs are selected, the input scaling and the number of selected nodes from the ESN reservoir (see Section 6.1).…”

“…However, this new work further extends the study of the influence of other parameters on the quality of the segmentation by including: the effect of the selected color space, the order in which inputs are selected, the input scaling and the number of selected nodes from the ESN reservoir (see Section 6.1). Furthermore, while the previous study 49 used only one dataset, namely, the semantic segmentation dataset (SSDS), 51 in the present work we have conducted a wide range of experiments using popular datasets such as the segmentation evaluation dataset (SED) 52 and the DRIVE 39 datasets. In Reference 50, the ESN‐based framework has been applied for segmenting blood vessels in retinal images using the DRIVE dataset where 13 pixel features were used (the RGB chromatic values, the mean, the SD, the gradient magnitude and angle in addition to six Gabor features).…”

Reservoir Computing Approach for Gray Images Segmentation

“…Similar works dealing with ESN features extraction for image segmentation were reported in [12,17,18,19]. Other works propose to train ESN to classify image pixels based on their preliminary extracted features [2,8,9,10,13,14].…”

Optimal modularity and memory capacity of neural reservoirs

“…Compared with other frameworks that require training numerous parameters, this paradigm allows for larger networks and better parameter scaling. Reservoir computers have been successful in a range of tasks including time series prediction, natural language processing, and pattern generation, and have also been used as biologically plausible models for neural computation (Deng, Mao, & Chen, 2016; Enel et al, 2016; Holzmann & Hauser, 2010; Jaeger, 2012; Jalalvand, De Neve, Van de Walle, & Martens, 2016; Rössert, Dean, & Porrill, 2015; Soriano et al, 2015; Souahlia, Belatreche, Benyettou, & Curran, 2016; Triefenbach, Jalalvand, Schrauwen, & Martens, 2010; Yamazaki & Tanaka, 2007).…”