Toward a Reliable Collection of Eye-Tracking Data for Image Quality Research: Challenges, Solutions, and Applications

Comparison of Visual Saliency for Dynamic Point Clouds: Task-free vs. Task-dependent

This paper presents a Task-Free eye-tracking dataset for Dynamic Point Clouds (TF-DPC) aimed at investigating visual attention. The dataset is composed of eye gaze and head movements collected from 24 participants observing 19 scanned dynamic point clouds in a Virtual Reality (VR) environment with 6 degrees of freedom. We compare the visual saliency maps generated from this dataset with those from a prior task-dependent experiment (focused on quality assessment) to explore how high-level tasks influence human visual attention. To measure the similarity between these visual saliency maps, we apply the well-known Pearson correlation coefficient and an adapted version of the Earth Mover's Distance metric, which takes into account both spatial information and the degrees of saliency. Our experimental results provide both qualitative and quantitative insights, revealing significant differences in visual attention due to task influence. This work enhances our understanding of the visual attention for dynamic point cloud (specifically human figures) in VR from gaze and human movement trajectories, and highlights the impact of task-dependent factors, offering valuable guidance for advancing visual saliency models and improving VR perception.

Clear Aligners and Smart Eye Tracking Technology as a New Communication Strategy between Ethical and Legal Issues

Smart eye-tracking technology (SEET) that determines visual attention using smartphones can be used to determine the aesthetic perception of different types of clear aligners. Its value as a communication and comprehension tool, in addition to the ethical and legal concerns which it entails, can be assessed. One hundred subjects (50 F, 50 M; age range 15-70) were equally distributed in non-orthodontic (A) and orthodontic (B) groups. A smartphone-based SEET app assessed their knowledge of and opinions on aligners. Subjects evaluated images of smiles not wearing aligners, with/without attachments and with straight/scalloped gingival margins, as a guided calibration step which formed the image control group. Subsequently, the subjects rated the same smiles, this time wearing aligners (experimental images group). Questionnaire data and average values for each group of patients, and images relating to fixation times and overall star scores, were analyzed using these tests: chi-square, t-test, Mann-Whitney U, Spearman's rho, and Wilcoxon (p < 0.05). One-way ANOVA and related post-hoc tests were also applied. Orthodontic patients were found to be better informed than non-orthodontic patients. Aesthetic perception could be swayed by several factors. Attachments scored lower in aesthetic evaluation. Lips distracted attention from attachments and improved evaluations. Attachment-free aligners were better rated overall. A more thorough understanding as to the opinions, expectations and aesthetic perception of aligners can improve communication with patients. Mobile SEET is remarkably promising, although it does require a careful medicolegal risk-benefit assessments for responsible and professional use.

Face emotion recognition based on infrared thermal imagery by applying machine learning and parallelism

Over time for the past few years, facial expression identification has been a promising area. However, darkness, lighting conditions, and other factors make facial emotion identification challenging to detect. As a result, thermal images are suggested as a solution to such problems and for a variety of other benefits. Furthermore, focusing on significant regions of a face rather than the entire face is sufficient for reducing processing and improving accuracy at the same time. This research introduces novel infrared thermal image-based approaches for facial emotion recognition. First, the entire image of the face is separated into four pieces. Then, we accepted only four active regions (ARs) to prepare training and testing datasets. These four ARs are the left eye, right eye, and lips areas. In addition, ten-folded cross-validation is proposed to improve recognition accuracy using Convolutional Neural Network (CNN), a machine learning technique. Furthermore, we incorporated a parallelism technique to reduce processing-time in testing and training datasets. As a result, we have seen that the processing time reduces to 50%. Finally, a decision-level fusion is applied to improve the recognition accuracy. As a result, the proposed technique achieves a recognition accuracy of 96.87 %. The achieved accuracy ascertains the robustness of our proposed scheme.

Pandemic disease detection through wireless communication using infrared image based on deep learning

Rapid diagnosis to test diseases, such as COVID-19, is a significant issue. It is a routine virus test in a reverse transcriptase-polymerase chain reaction. However, a test like this takes longer to complete because it follows the serial testing method, and there is a high chance of a false-negative ratio (FNR). Moreover, there arises a deficiency of R.T.-PCR test kits. Therefore, alternative procedures for a quick and accurate diagnosis of patients are urgently needed to deal with these pandemics. The infrared image is self-sufficient for detecting these diseases by measuring the temperature at the initial stage. C.T. scans and other pathological tests are valuable aspects of evaluating a patient with a suspected pandemic infection. However, a patient's radiological findings may not be identified initially. Therefore, we have included an Artificial Intelligence (A.I.) algorithm-based Machine Intelligence (MI) system in this proposal to combine C.T. scan findings with all other tests, symptoms, and history to quickly diagnose a patient with a positive symptom of current and future pandemic diseases. Initially, the system will collect information by an infrared camera of the patient's facial regions to measure temperature, keep it as a record, and complete further actions. We divided the face into eight classes and twelve regions for temperature measurement. A database named patient-info-mask is maintained. While collecting sample data, we incorporate a wireless network using a cloudlets server to make processing more accessible with minimal infrastructure. The system will use deep learning approaches. We propose convolution neural networks (CNN) to cross-verify the collected data. For better results, we incorporated tenfold cross-verification into the synthesis method. As a result, our new way of estimating became more accurate and efficient. We achieved 3.29% greater accuracy by incorporating the "decision tree level synthesis method" and "ten-folded-validation method". It proves the robustness of our proposed method.

How is Gaze Influenced by Image Transformations? Dataset and Model

Data size is the bottleneck for developing deep saliency models, because collecting eye-movement data is very time-consuming and expensive. Most of current studies on human attention and saliency modeling have used high-quality stereotype stimuli. In real world, however, captured images undergo various types of transformations. Can we use these transformations to augment existing saliency datasets? Here, we first create a novel saliency dataset including fixations of 10 observers over 1900 images degraded by 19 types of transformations. Second, by analyzing eye movements, we find that observers look at different locations over transformed versus original images. Third, we utilize the new data over transformed images, called data augmentation transformation (DAT), to train deep saliency models. We find that label-preserving DATs with negligible impact on human gaze boost saliency prediction, whereas some other DATs that severely impact human gaze degrade the performance. These label-preserving valid augmentation transformations provide a solution to enlarge existing saliency datasets. Finally, we introduce a novel saliency model based on generative adversarial networks (dubbed GazeGAN). A modified U-Net is utilized as the generator of the GazeGAN, which combines classic "skip connection" with a novel "center-surround connection" (CSC) module. Our proposed CSC module mitigates trivial artifacts while emphasizing semantic salient regions, and increases model nonlinearity, thus demonstrating better robustness against transformations. Extensive experiments and comparisons indicate that GazeGAN achieves state-of-the-art performance over multiple datasets. We also provide a comprehensive comparison of 22 saliency models on various transformed scenes, which contributes a new robustness benchmark to saliency community. Our code and dataset are available at.