Accurate object recognition in the underwater images using learning algorithms and texture features

Currently clustering techniques play a vital role in object recognition process. The clustering techniques are found to be efficient with neural networks. So, the present paper proposed a novel method for clustering the input objects with Self-Organizing Map (SOM). The proposed method considers the input object as a random closed set. The random set can be efficiently described with various features viz., volume fractions, covariance and contact distributions etc. In the proposed method, the input object is described efficiently with spherical contact distribution. The proposed method is experimented with the leaf data set with 795 images. The performance of the proposed method is evaluated with various topologies of SOM and is measured with four measures viz., FNR, FPR, TPR and TNR. The results indicate the efficiency of the proposed method.

show abstract

“…The Spherical contact distribution function (SCDF) of S with a ball of radius t centered at origin is given in (3).…”

Section: Methodsmentioning

confidence: 99%

“…It consists of three stages viz., background subtraction, feature extraction, and training and classification. The underwater recognition system [3] is designed with chain coding, texture and statistical features. The described feature vector is given to a learning algorithm for getting efficient result.…”

Section: Introductionmentioning

confidence: 99%

Contact Distribution Function based Clustering Technique with Self-Organizing Maps

Chamundeswari¹,

Varma²,

Satyanarayana³

2018

IJIGSP

View full text Add to dashboard Cite

show abstract

“…Additionally, the problem is more compounded due to poor illumination conditions. Srividhya et al [48] have extracted different statistical features like autocorrelation and sum of entropy which are used by the learning algorithms to classify underwater objects. To detect underwater moving object, Liu et al [21] have proposed a underwater object detection scheme which combines the notions of background subtraction and three frame differences under the assumption of a fixed camera position.…”

Section: Related Workmentioning

confidence: 99%

Kernel density estimation and correntropy based background modeling and camera model parameter estimation for underwater video object detection

Panda

Nanda

2021

Soft Comput

View full text Add to dashboard Cite

Underwater video object detection is challenging because of the complex background and the movement of the camera. In order to address this, we propose a novel scheme of simultaneously estimating the camera model parameters and detecting the object. The object detection phase includes background modeling and its learning. Background is modeled by the proposed Spatial Kernel Density Estimation (SKDE) model and the model learning happens in the SKDE feature space. Background modeling and its learning is pixel based approach. The model histograms learn the new pixel through its histogram representation. Our learning and classification strategy is different from the Heikkila et al. [17] in the context of similarity measure. We have proposed the correntropy based similarity measure that is used for model learning and pixel classification. The camera model parameters are estimated by 2D optimization method where we have used the corner features of an object at subpixel accuracy level. These subpixel level features are used in the proposed pipelining framework for model parameters esti-

show abstract

“…ROSUB 6000 was able to capture rare underwater coral, fishes, worm-like species, as oxygen levels and temperature is very low (oxygen less than 0.5 mg/l and temperature 5 °C) at such depth in water. 12 images were collected from NIOT, 120 images from fish4knowledge and 120 images from Fishdb website, out of which 152 images were used for training and 100 images were used for testing [26].…”

Section: The First Examined Datamentioning

confidence: 99%

A joint adaptive evolutionary model towards optical image contrast enhancement and geometrical reconstruction approach in underwater remote sensing

Moghimi

Mohanna

2019

SN Appl. Sci.

View full text Add to dashboard Cite

Nowadays, the underwater optical imaging is used by a wide range of sciences including marine scientists for studying underwater structures and organisms, oil and gas companies, telecommunications to monitor fuel transmission lines and cables, and military centers to detect the sea mines and submarines. Images with non-uniform brightness are visually limited due to low ambient light during imaging or excessive exposure, fogging, hazing, or a combination of these factors. As a result, a process of imaging improvement is highly complex with some conditions. Particularly, this issue is harmful for the underwater environments as well as its highly complexity. In this paper, we improve the underwater images from different environments by exploiting image processing techniques including elimination of fogging in the image, and improvement of contrast with combining super resolution methods which result in significant enhancement of the image quality. To remove fogging in the image enhancement method, a variational-based fusion method was proposed without increasing or decreasing image resolution to increase the adaptability of non-uniform backlight images. This approach is optimized by the bee colony algorithm. Additionally, weighting coefficient selection method is used for super resolution of image after adjusting contrast. The main reason for using hybrid contrast techniques includes the adaptive enhancement and color correction. The bee colony algorithm and weighting coefficient in super resolution due to the lack of hybrid methods in visually improved water images has been the subject of recent researches. The hybrid set has provided more effect on different images at various conditions. Having applied the algorithm set for low-quality visual image set of underwater, some analyses such as PSNR, MSE and SSIM are tested. Compared to similar solutions, the quality is significantly improved and highlighted the exhaustive operation of the algorithm. Applying the presented algorithm in the software part of underwater imaging systems is widely effective for the accurate analysis of underwater images.

show abstract

Accurate object recognition in the underwater images using learning algorithms and texture features

Cited by 19 publications

References 18 publications

Contact Distribution Function based Clustering Technique with Self-Organizing Maps

Contact Distribution Function based Clustering Technique with Self-Organizing Maps

Kernel density estimation and correntropy based background modeling and camera model parameter estimation for underwater video object detection

A joint adaptive evolutionary model towards optical image contrast enhancement and geometrical reconstruction approach in underwater remote sensing

Contact Info

Product

Resources

About