No-reference image quality assessment algorithms: A survey

In the field of image processing, there are several problems where an efficient search of the solutions has to be performed within a complex search domain to find an optimal solution. Multi-thresholding which is a very important image segmentation technique is one of them. The multi-thresholding problem is simply an exponential combinatorial optimization process which traditionally is formulated based on complex objective function criterion which can be solved using only nondeterministic methods. Under such circumstances, there is also no unique measurement which quantitatively judges the quality of a given segmented image. Therefore, researchers are solving those issues by using Nature-Inspired Optimization Algorithms (NIOAs) as alternative methodologies for the multi-thresholding problem. This study presents an up-to-date review on all most important NIOAs employed in multi-thresholding based image segmentation domain. The key issues which are involved during the formulation of NIOAs based image multi-thresholding models are also discussed here.

Nature-Inspired Optimization Algorithms and Their Application in Multi-Thresholding Image Segmentation

Local gradient patterns (LGP): An effective local-statistical-feature extraction scheme for no-reference image quality assessment

Early detection of induction-motor faults has been an increasing matter of research in the last few years. The reliable identification of broken rotor bars (BRB) is still under investigation as it is one of the most common and difficult-to-detect faults in induction motors. Many methods have been proposed to deal with this issue. Recent approaches combine techniques looking for improving the performance of the diagnosis. Their major disadvantage is the high computational requirements, which restrains them from being used in online detection. The contribution in this paper is twofold. The first one is a novel methodology for induction motor BRB detection and the fault severity classification using homogeneity as index, which, to the best of our knowledge, has never been used as an indicator for fault diagnosis, analyzing one phase of the induction motor startup-transient current. Because of the low computational complexity in homogeneity calculation, the second contribution of this paper is a hardware-processing unit based on a field programmable gate array device for online detection and classification of BRB. Obtained results demonstrate the high efficiency of the proposed methodology as a deterministic technique for incipient BRB diagnosis in induction motors, which can detect and differentiate among half, one, or two BRBs with a certainty greater than 99.7%.

Novel FPGA-based Methodology for Early Broken Rotor Bar Detection and Classification Through Homogeneity Estimation

Edge-preserving image denoising using a deep convolutional neural network

An image despeckling approach using quantum-inspired statistics in dual-tree complex wavelet domain

Saliency computational model with active environment perception can be useful for many applications including image segmentation, image compression, image retrieval, and etc. Conventional saliency computational models rely on handcrafted low level features, such as color or contrast. These models face great difficulties in low lighting scenarios, due to the lack of well-defined feature to interpret saliency information in low contrast images. In this paper, a new approach is proposed to detect salient object from low contrast images. The proposed approach explores the most distinguishable salient information in low contrast images based on low level features. Extensive experiments have been conducted to evaluate the performance of the proposed method against the state-of-the-art saliency computational models.

Salient object detection from distinctive features in low contrast images

A novel medical ultrasound despeckling method combining quantum-inspired bilateral filtering and wavelet thresholding is proposed. An adaptive bilateral filter under the framework of quantum signal processing is exploited in this method and used as a pre-processing step. Then wavelet thresholding, on the assumption that the signal and noise coefficients in the wavelet domain are subject to generalised Laplace distribution and Gaussian distribution, respectively, is applied to the pre-filtered image. Experiments show that the proposed method can obtain a better speckle noise reduction performance while retaining high-contrast features such as edges.

Quantum-inspired hybrid medical ultrasound images despeckling method

Blind noisy image quality assessment using block homogeneity

The objective of blind noisy image quality assessment is to evaluate the quality of the degraded noisy image without the knowledge of the ground truth image. Its performance relies on the accuracy of the noise statistics estimated from homogenous blocks. The major challenge of block-based approaches lies in the block size selection, as it affects the local noise derivation. To tackle this challenge, a deformable ant colony optimization (DACO) approach is proposed in this paper to adaptively adjust the ant size for image block selection. The proposed DACO approach considers that the size of the ant is adjustable during foraging. For the smooth image blocks, more pheromone is deposited, and then the size of ant is increased. Therefore, this strategy enables the ants to have dynamic food-search capability, leading to more accurate selection of homogeneous blocks. Furthermore, the regression analysis is used to obtain image quality score by exploiting the above-estimated noise statistics. Experimental results are provided to justify that the proposed approach outperforms conventional approaches to provide more accurate noise statistics estimation and achieve a consistent image quality evaluation performance for both the artificially generated and real-world noisy images.

Xiaowei Fu

Papers

An image despeckling approach using quantum-inspired statistics in dual-tree complex wavelet domain

Salient object detection from distinctive features in low contrast images

Quantum-inspired hybrid medical ultrasound images despeckling method

Blind noisy image quality assessment using block homogeneity

Blind noisy image quality evaluation using a deformable ant colony algorithm