No-reference quality assessment for image-based assessment of economically important tropical woods

Abstract

Image Quality Assessment (IQA) is essential for the accuracy of systems for automatic recognition of tree species for wood samples. In this study, a No-Reference IQA (NR-IQA), wood NR-IQA (WNR-IQA) metric was proposed to assess the quality of wood images. Support Vector Regression (SVR) was trained using Generalized Gaussian Distribution (GGD) and Asymmetric Generalized Gaussian Distribution (AGGD) features, which were measured for wood images. Meanwhile, the Mean Opinion Score (MOS) was obtained from the subjective evaluation. This was followed by a comparison between the proposed IQA metric, WNR-IQA, and three established NR-IQA metrics, namely Blind/Referenceless Image Spatial Quality Evaluator (BRISQUE), deepIQA, Deep Bilinear Convolutional Neural Networks (DB-CNN), and five Full Reference-IQA (FR-IQA) metrics known as MSSIM, SSIM, FSIM, IWSSIM, and GMSD. The proposed WNR-IQA metric, BRISQUE, deepIQA, DB-CNN, and FR-IQAs were then compared with MOS values to evaluate the performance of the automatic IQA metrics. As a result, the WNR-IQA metric exhibited a higher performance compared to BRISQUE, deepIQA, DB-CNN, and FR-IQA metrics. Highest quality images may not be routinely available due to logistic factors, such as dust, poor illumination, and hot environment present in the timber industry. Moreover, motion blur could occur due to the relative motion between the camera and the wood slice. Therefore, the advantage of WNR-IQA could be seen from its independency from a "perfect" reference image for the image quality evaluation.

Are Convolutional Neural Networks Trained on ImageNet Images Wearing Rose‐Colored Glasses?: A Quantitative Comparison of ImageNet, Computed Tomographic, Magnetic Resonance, Chest X‐Ray, and Point‐of‐Care Ultrasound Images for Quality

“…Prior work with BRISQUE and CNN development has shown improved performance via its superior ability to evaluate CNN training images for noise, distortion, and quality compared to contemporary image quality assessment measures. 16 Additionally, CNN developers have successfully used BRISQUE to improve the quality of training and test images by using it to denoise and improve lower-quality images. 17 Although most CNN development and training has taken place on nonmedical images, there is evidence that poor-quality medical images negatively affect CNN performance for medical diagnosis as well.…”

Development of Image Quality Assessment (IQA) For Haze Prediction

“…Firstly, the Mean Subtracted Contrast Normalized (MSCN) of the hazy images were calculated [6]. Then, two types of Gaussian distribution functions were incorporated in this study to accommodate the diverse characteristics of MSCN coefficient, namely the Generalized Gaussian Distribution (GGD) and Asymmetric Generalized Gaussian Distribution (AGGD) [5] . The GGD, where α represents the shape of the distribution and σ 2 represents the variance and AGGD parameters were calculated using the Eqs.…”

“…The IQA is classified into two methods, namely, subjective and objective methods [4]. The subjective method is used to obtain human's perception on the image quality [5]. This method simulates the perception technique of a people, a visual system to estimate it.…”

A No-Reference Image Quality Assessment Metric for Wood Images

“…The MSCN coefficients, I m n  ( , ) highlights the main features of the wood images such as pores and grains, with few low-energy residual object boundaries [21]. Therefore, the MSCN is used to compute the GLCM and Gabor features instead of the image, I(m,n).…”

“…FR-IQA uses the reference image fully to assess the images [10][11][12][13][14][15] whereas RR-IQA uses the reference images partially [16,17]. In contrast, NR-IQA assesses an image without using a reference image [18][19][20][21]. NR-IQA is the most appropriate algorithm to evaluate the quality of the wood images as it may be impossible to obtain high quality images in the dusty and dark setting of lumber factories.…”