Chouhan, Rajlaxmi

Distinguishing a natural scene from artwork is a simple task for the human visual system, but a challenging one for machines due to the wide range of psychovisual features, illumination gamuts, and varied interpretations of glossiness. While state-of-the-art image quality metrics quantify overall visual quality to a remarkable degree of accuracy, quantification of ‘glossiness’ of a scene to represent its naturalness is still an emerging area. The rapid growth of deep learning methods and CNN-based architectures inspired us to explore the fusion of best performing CNN architectures in this paper for an image dataset specifically created to replicate unnaturalness of artwork or portrait-like images. Performance of four CNN networks was tested using transfer learning, selective retraining and optimizing initial learning rates on a dataset of about 8.5k images created to represent various degrees of glossiness. A fusion framework was then proposed using the top two architectures. In terms of eleven levels of naturalness (0 to 10), both quantitative and qualitative evaluation of the fusion frameworks was conducted. The framework resulting from fusion of GoogleNet and VGG16, referred to as GoogleVGG Fusion in this paper, is found to reach accuracies comparable to individual networks but with nearly half the computational cost. The proposed GoogleVGG Fusion model achieved an accuracy of 87.86% with the labelled scores and a Spearman’s Rank Correlation (SROCC) of 0.9794. As expected, the accuracy of the proposed framework with the subjective scores in comparison with non–deep learning (DL) & DL-based methods is remarkably better.

This paper presents simple and intuitive strategies for effective online assessment of a freshman engineering course. The proposed strategies for unproctored online exams include creating multiple sets with identical options, using image-format questions, maintaining short duration of exam, and employing a rotational assignment. Unlike a single-set paper where the class average was found to be disproportionately high, the proposed strategies helped in correctly restoring the class average. Feedback and results from 141 students show a significant statistical difference in the scores obtained using the proposed multi-set quiz framework, as opposed to conventional single-set papers. The paper also presents a simple setup of remote proctoring using dual-video and screensharing to replicate in-person classroom exams without the need of commercial proctoring services.

Preserving edges while denoising an image is a crucial and challenging necessity. In this paper, a noise-aided edge-preserving denoising algorithm is proposed by extending the classical patch-based Non Local Means (NLM) algorithm. The classical NLM algorithm uses similarity weights to ensure edge preservation. In the proposed algorithm, these similarity weights are enhanced by processing the similarity distances using stochastic resonance (SR). SR refers to a phenomenon where the performance metric in a nonlinear system counterintuitively increases to a peak and then decreases (like a bell curve) in the presence of a controlled amount of noise. The similarity weights derived from NLM are iteratively processed using the discretized SR equation. For the iteratively reconstructed images, the local maxima of the corresponding quality metric, PSNR, is selected as the optimal output. The iterative processing results in a nonlinear scaling of the similarity distances. This processing, in effect, ensures that similarity weights of similar patches are high and those of dissimilar patches are low, thereby producing enhanced edge preservation. The performance of the proposed algorithm is demonstrated by presenting the comparative results for a variety of images corrupted by a wide range of AWGN noise. The proposed algorithm is found to handle the spurious artifacts near the edges more efficiently. Moreover, the obtained edges are sharper and better preserved even in the presence of high noise deviation. Benchmarking results on SET12 and BSD68 datasets show an improvement of 14.5% and 12.1% respectively over that of NLM for high noise.

This paper aims to present an analysis of a noise-aided contrast enhancement algorithm in hybrid transform domains. The performance of our earlier noise-enhanced iterative algorithm, formulated from the motion dynamics of a double-well system exhibiting dynamic stochastic resonance, has been investigated here on hybrid coefficients, viz. singular values (SVs) of wavelet coefficients, SVs of discrete cosine transform (DCT) coefficients, and DCT of wavelet coefficients, of a dark image. The performance of the algorithm is gauged using metrics indicating relative contrast enhancement and perceptual quality. Colorfulness, subjective visual scores and logarithmic contrast metrics for outputs are also observed. Experimental results display noteworthy enhancement of contrast on both natural and synthetically-darkened images. It can be inferred from comparative analysis with respect to other conventional methods that while the algorithm is observed to work well in all three hybrid domains, the SV-DCT domain performs better in terms of iteration count, while DCT-DWT is found to outperform others in terms of perceptual quality.

Image quality assessment (IQA) plays a crucial role in monitoring quality control in image communication systems, and in benchmarking and optimizing parameters in enhancement algorithms. The full-reference IQA metrics require a good-quality reference image, obtaining which may not be practical in real-life applications. This paper, therefore, proposes a no-reference IQA metric based on the hypothesis that every image has latent additive white Gaussian noise (AWGN). A mathematical model was developed on a dataset of fifty test images by computing gradient-based structural similarity of corrupted images w.r.t. the original. Statistical modeling of the observations were found to fit an exponential parametric model. The standard deviation of the latent (or apparent) AWGN present in any image was estimated using an SVD-based approach. The proposed metric, referred to as the no-reference gradient-based structural integrity (NRGSI), is then computed by a simple backprojection of the estimated noise deviation on the exponential model. The accuracy of the proposed objective metric is characterized by its comparison with subjective quality scores given by ten subjects, and with a classical perceptual quality measure.

In the world of ever-growing technology and multimedia devices, educators around the globe are innovating newer ways to engage students in a more immersive and enjoyable way. Games have had a proven effect on marketing, sales, healthcare, behavioral changes, K12 education, and have now increasingly gained momentum in higher education, especially engineering education, primarily to enhance learner engagement and motivation. This paper presents a work-in-progress towards developing serious games as supplemental material in teaching the course on Basic Electronics. Two story boards of game-based learning are presented in this paper along with the design strategy and target core motivational drives. The game development is under progress and student reception and feedback is yet to be incorporated.

The use of smartphone cameras for capturing photographs has seen an exponential growth in the last decade. The noise present in these photographs significantly deviates from the popular Independently Identically Distributed (i.i.d.) Additive White Gaussian Noise (AWGN) noise, and thus the conclusions drawn from simulated-AWGN cannot be directly applied to the smartphone's true noise. This paper is the first reporting of the exhibition of Stochastic Resonance (SR) or noise-induced threshold-crossing in three genres of discontinuity detectors used in image processing - corner detector, line detector, and edge detector for real-world smartphone images. For the images under investigation, the performance of these detectors is quantified w.r.t. parameters representing intrinsic noise. Observations suggest that all these detectors inherently exhibit the phenomenon of SR due to the fundamental assistance offered by controlled amount of noise in crossing detector thresholds. The manifestations of SR - constant parameter value with varying noise and varying parameter value with constant noise - are demonstrated to exhibit SR in each of the three detectors.

In this paper, a preliminary case study of gauging effectiveness of educational videos is conducted on a sample video created for freshman batch of novice electrical engineering students. The video was assigned in flipped classroom mode, a learner-centered pedagogical model, as a video assignment to students with five interleaved knowledge check questions. The 'engagement' of the student with the video material can be gauged in terms of both watchtime and number of times a particular section of the video was rewatched, along with the obtained score. The proposed hypothesis is that the placement of knowledge check questions along with the complexity of the topic covered together can be correlated with the number of times a section of the video is, or needs to be, rewatched by the student. Observations for 25 novice students are correlated with ground truth of complexity proposed by the instructor, and suggests that in addition to the knowledge checkpoints and complexity, elements of storytelling in pedagogy also play an interesting role in determining engagement of the learner with the video material.

Blended learning using flipped classroom has emerged as a remarkable model for bridging the digital divide between geographically- and socio-economically-disadvantaged students during the Covid pandemic, especially for those situated in remote locations with limited internet connectivity. This paper presents an investigation of a blended learning framework using flipped teaching through interactive video lectures for an Introduction to Electrical Engineering course held in online mode for a first-year batch of engineering students. The analysis includes observations from two sets of students who took this course over the last two academic years, referred to as control and treatment groups, respectively. The paper also reports a simple preliminary framework for utilizing the analytics available from interactive video assignments for identifying students with poor engagement and understanding so that the instructor may adopt timely measures and interventions to address the need of all students. The success of the video assignments are validated both qualitatively (through student feedback and performance) and quantitatively (using t-test). Results show that the flipped teaching model has both empirical approval and support of a majority of students, especially in the distance learning mode. Statistical analysis shows a significant difference in the performance of students with and without video-based flipped teaching in online mode.

Noise-aided stochastic resonance has been explored in recent literature as a powerful tool that enhances the performance of non-linear systems, particularly in image enhancement and image watermarking. In this paper, we extend the application of stochastic resonance to improve the performance of the conventional non-local means (NLM) filtering for edge-preserving image denoising. The NLM algorithm typically involves computation of weights denoting similarity of a pixel with all other pixels in the image. In the proposed algorithm, these similarity weights are iteratively processed using the concept of dynamic stochastic resonance. The results indicate a significant improvement in sharpness of edges in the denoised images in comparison with the conventional NLM approach both visually and quantitatively in terms of full-reference and no-reference image quality metrics.