Perceptual Measures Research Articles

Recovery From Eccentric Squat Exercise in Resistance-Trained Young and Master Athletes With Similar Maximum Strength: Combining Cold Water Immersion and Compression.

The aim of this study was to investigate whether recovery from eccentric squat exercise varies depending on age and to assess whether the use of a mixed-method recovery (MMR) consisting of cold water immersion and compression tights benefits recovery. Sixteen healthy and resistance-trained young (age, 22.1±2.1years; N=8) and master male athletes (age, 52.4±3.5years; N=8), who had a similar half squat 1-repetition maximum relative to body weight, completed two identical squat exercise training sessions, separated by a 2-week washout period. Training sessions were followed by either MMR or passive recovery (PR). Internal training loads [heart rate and blood lactate concentration (BLa)] were recorded during and after squat sessions. Furthermore, maximal voluntary isometric contraction (MVIC) force, countermovement jump (CMJ) height, resting twitch force of the knee extensors, serum concentration of creatine kinase (CK), muscle soreness (MS), and perceived physical performance capability (PPC) were determined before and after training as well as after 24, 48, and 72h of recovery. A three-way mixed ANOVA revealed a significant time effect of the squat protocol on markers of fatigue and recovery (p<0.05; decreased MVIC, CMJ, twitch force, and PPC; increased CK and MS). Age-related differences were found for BLa, MS, and PPC (higher post-exercise fatigue in younger athletes). A significant two-way interaction between recovery strategy and time of measurement was found for MS and PPC (p<0.05; faster recovery after MMR). In three participants (two young and one master athlete), the individual results revealed a consistently positive response to MMR. In conclusion, master athletes neither reach higher fatigue levels nor recover more slowly than the younger athletes. Furthermore, the results indicate that MMR after resistance exercise does not contribute to a faster recovery of physical performance, neuromuscular function, or muscle damage, but promotes recovery of perceptual measures regardless of age.

Analysis of Early Biomarkers Associated With Noise-Induced Hearing Loss Among Shipyard Workers

It is important to determine what frequencies and auditory perceptual measures are the most sensitive early indicators of noise-induced hearing impairment. To examine whether hearing loss among shipyard workers increases more rapidly at extended high frequencies than at clinical frequencies and whether subtle auditory processing deficits are present in those with extensive noise exposure but little or no hearing loss. This cross-sectional study collected audiometric data (0.25-16 kHz), survey questionnaires, and noise exposure levels from 7890 shipyard workers in a Shanghai shipyard from 2015 to 2019. Worsening hearing loss was evaluated in the group with hearing loss. Speech processing and temporal processing were evaluated in 610 participants with noise exposure and clinically normal hearing to identify early biomarkers of noise-induced hearing impairment. Data analysis was conducted from November to December 2020. Linear regression was performed to model the increase in hearing loss as function of cumulative noise exposure and compared with a group who were monitored longitudinally for 4 years. Auditory processing tests included speech-in-noise tests, competing sentence tests, dichotic listening tests, and gap detection threshold tests and were compared with a control group without history of noise exposure. Of the 5539 participants (median [interquartile range (IQR)] age, 41.0 [34.0-47.0] years; 3861 [86.6%] men) included in the cross-sectional analysis, 4459 (80.5%) were hearing loss positive and 1080 (19.5%) were hearing loss negative. In younger participants (ie, ≤40 years), the maximum rate of increase in hearing loss was 0.40 (95% CI, 0.39-0.42) dB per A-weighted dB-year (dB/dBA-year) at 12.5 kHz, higher than the growth rates of 0.36 (95% CI, 0.35-0.36) dB/dBA-year at 4 kHz, 0.32 (95% CI, 0.31-0.33) dB/dBA-year at 10 kHz, 0.31 (95% CI, 0.30-0.31) dB/dBA-year at 6 kHz, 0.27 (95% CI, 0.26-0.27) dB/dBA-year at 3 kHz, and 0.27 (95% CI, 0.27-0.28) dB/dBA-year at 8 kHz. In the 4-year longitudinal analysis of hearing loss among 403 participants, the mean (SD) annual deterioration in hearing was 2.70 (2.98) dB/y at 12.5 kHz, almost twice as that observed at lower frequencies (eg, at 3kHz: 1.18 [2.15] dB/y). The auditory processing scores of participants with clinically normal hearing and a history of noise exposure were significantly lower than those of control participants (eg, median [IQR] score on speech-in-noise test, noise-exposed group 1 vs control group: 0.63 [0.55-0.66] vs 0.78 [0.76-0.80]; P < .001). These findings suggest that the increase in hearing loss among shipyard workers was more rapid at 12.5 kHz than at other frequencies; workers with clinically normal hearing but high cumulative noise exposure are likely to exhibit deficits in speech and temporal processing.

The role of L2 experience in L1 and L2 perception and production of voiceless stops by English learners of Spanish

Some previous studies report that increased experience with a second language (L2) may result in a more target-like perception and production in the L2, as well as in a less native-like performance in the L1. The present paper aimed to (1) assess the role of L2 experience on L2 and L1 production of voiceless stops; (2) investigate the effect of L2 experience on L2 and L1 perception of voiceless stops; and (3) examine the relationship between perception and production. Three groups of English learners of Spanish differing in amount and type of L2 experience, as well as two groups of functional monolinguals, completed a production task and an identification task involving English and Spanish voiceless stops. The results revealed that the L2 speakers were more successful at producing than at perceiving Spanish stops accurately, with L2 experience having a positive effect on production. L2 experience was not found to affect performance in the L1, which could be related to an overall limited amount of L2 use even in an immersion setting. The results also showed a weak relationship between perception and production, which may partly be due to the different nature of perceptual and production measures.

Intrinsic value of food chain data

Intrinsic value of food chain data

Effects of Traffic Control Devices on Rural Curve Lateral Position

This study investigated, by means of a dynamic driving simulator experiment, road users’ behavior inside the curves of rural two-lane highways related to different advance warning signs, perceptual measures, and delineation treatments. These treatments were intended to warn drivers of the presence of low radius curves and to affect their behavior, improving vehicle control and lane-keeping. Five surrogate measures of safety were used in the evaluation of the design alternatives in relation to lateral position performance: standard deviation of lateral position, maximum encroachment in the shoulder, maximum encroachment in the opposite lane, relative length of shoulder encroachment, and relative length of opposite lane encroachment. Statistical tests were performed to verify whether the surrogate measures of safety were significantly different between alternatives. The analysis was divided into three phases. In the first phase, all the measures were preprocessed, testing normality and homoscedasticity assumptions. In the second phase, the presence of an overall effect considering all the alternatives was evaluated using analysis of variance and the Kruskal–Wallis test. In the third phase, Student’s t- and Mann–Whitney tests were used to assess which alternatives showed statistically significant effects. The results demonstrated that the perceptual measures, namely colored transverse strips, dragon teeth markings, and colored median island, were the most effective treatments. Field tests to establish the perceptual measures’ effectiveness on real roads are strongly advised. Implementation of the measures tested in the driving simulator should be carried out on similar rural highways to validate the results.

Noise perception assessment in urban green spaces using soundwalk approach

Urban green spaces are intended to provide citizens with calm environments free of annoying city noises. This requires a thorough understanding of noise emission and related exposure to sounds in green spaces. This research investigates noise perception in various spots in an urban green space. For this purpose, the study has been conducted in the grand garden of the city of Dresden. The garden covers 1.8 square kilometers of various landscapes, including water streams, park railways, fountains, bridges, roads for bicycles and pedestrians etc. Noise perception was investigated at eleven spots with emphasis on four noise types: nature noise, human noise, traffic noise, and technical noise. In parallel, audio-visual recordings were conducted for each spot to identify the connection between the perceptual measures and the psychoacoustic parameters. These spots are categorized based on the resulting perception and psychoacoustic parameters. In addition, the visual effect of each spot on final perception is investigated. Eventually, annoyance for each spot is identified based on the corresponding participants' perception and is associated with the relevant psychoacoustic parameters.

Phonetic convergence to non-native speech: Acoustic and perceptual evidence

While the tendency of speakers to align their speech to that of others acoustic-phonetically has been widely studied among native speakers, very few studies have examined whether natives phonetically converge to non-native speakers. Here we measured native Dutch speakers’ convergence to a non-native speaker with an unfamiliar accent in a novel non-interactive task. Furthermore, we assessed the role of participants’ perceptions of the non-native accent in their tendency to converge. In addition to a perceptual measure (AXB ratings), we examined convergence on different acoustic dimensions (e.g., vowel spectra, fricative CoG, speech rate, overall f0) to determine what dimensions, if any, speakers converge to. We further combined these two types of measures to discover what dimensions weighed in raters’ judgments of convergence. The results reveal overall convergence to our non-native speaker, as indexed by both perceptual and acoustic measures. However, the ratings suggest the stronger participants rated the non-native accent to be, the less likely they were to converge. Our findings add to the growing body of evidence that natives can phonetically converge to non-native speech, even without any apparent socio-communicative motivation to do so. We argue that our results are hard to integrate with a purely social view of convergence.

Impact of Face Masks on Speech Acoustics and Vocal Effort in Healthcare Professionals.

We investigated speech acoustics and self-reported vocal symptoms in mask-wearing healthcare professionals. We hypothesized that there would be an attenuation of spectral energies and increase in vocal effort during masked speech compared to unmasked speech. Within and between subject quasi-experimental design. We prospectively enrolled 21 healthcare providers (13 cisgender female, 8 cisgender male; M=32.9 years; SD=7.9 years) and assessed acoustics and perceptual measures with and without a face mask in place. Measurements included: 1) acoustic Vowel Articulation Index (VAI); 2) cepstral and spectral acoustic measures; 3) traditional vocal measures (e.g., fundamental frequency, intensity); 4) relative fundamental frequency (RFF); and 5) self-reported ratings of vocal effort and dyspnea. During masked speech, there was a significant reduction in VAI, high-frequency information (>4 kHz), and RFF offset 10, as well as a significant increase in cepstral peak prominence and perceived vocal effort. Further analysis showed that high-frequency attenuation was more pronounced when wearing an N95 mask compared to a simple mask. Face masks pose an additional barrier to effective communication that primarily impacts spectral characteristics, vowel space measures, and vocal effort. Future work should evaluate how long-term mask use impacts vocal health and may contribute to vocal problems. 3 Laryngoscope, 132:391-397, 2022.

Improvements in Speech of Children with Apraxia: The Efficacy of Treatment for Establishing Motor Program Organization (TEMPOSM)

ABSTRACT Purpose: This study investigated the efficacy of Treatment for Establishing Motor Program Organization (TEMPOSM) in childhood apraxia of speech (CAS). Method: A mixed between- and within-participant design with multiple baselines across participants and behaviors was used to examine acquisition, generalization, and maintenance of skills. TEMPOSM was administered in four one-hour sessions a week over a four-week period for eleven participants (ages 5 to 8), allocated to either an immediate treatment group or a wait-list control group. Acoustic and perceptual variables were measured at baseline, immediate post-treatment, and one-month post-treatment. Results: Children demonstrated significant improvements in specific acoustic measures of segmentation and lexical stress, as well as perceptual measures of fluency, lexical stress, and speech-sound accuracy. Treatment and generalization effects were maintained one-month post-treatment with generalization to untreated stimuli. Conclusion: TEMPOSM was efficacious in improving segmental and suprasegmental impairments in the speech of children with CAS.

Perceptual similarity measurement based on generative adversarial neural networks in graphics design

Measuring the similarity between images is of paramount importance in computer vision. However, the commonly used pixelwise similarity metrics do not match well with perceptual similarity. The purpose of this paper is to propose a visual similarity measurement method, which can be effectively used for plagiarism detection in graphic design. Plagiarism detection of designs refers to the identification and determination of major similarities. It is difficult to carry out the similarity learning process in traditional deep neural network due to the insufficient of training samples. To overcome this problem, a novel scheme is proposed for measuring perceptual similarity of graphics by using a constraint Generative Adversarial Network (GAN) model. The generator of GAN is used to create similar graphics following the common plagiarism features of logo design. Unlike the traditional discriminator which judges the authenticity of the generated image and the original image, the modified discriminator is used to calculate the perceptual similarity of the graphics pair. In graphics design, plagiarism mainly focuses on the changes of shape, color and style, which has certain cognitive subjectivity. Therefore, design experts were invited to participate in a group of cognitive analysis experiments. A perceptual constraint model is established to limit the generation of plagiarized graphics according to “design and visual rationality”. Promising results demonstrate that the proposed method can be used for plagiarism detection of logo design. Given its effectiveness and conceptual simplicity, I hope it can serve as a baseline and contribute to the future research on plagiarism detection of artworks.

Voice Acoustics and Vocal Effort in Mask-Wearing Healthcare Professionals: A Comparison Pre- and Post-Workday

We evaluated voice acoustics and self-perceptual ratings in healthcare workers required to wear face masks throughout their workday. Eighteen subjects (11 cisgender female, 7 cisgender male; M=33.72 years, SD=8.30) completed self-perceptual ratings and acoustic recordings before and after a typical workday. Chosen measures were specific to vocal effort, dysphonia, and laryngeal tension. Mixed effects models were calculated to determine the impact of session, mask type, sex, and their interactions on the set of perceptual and acoustic measures. The subjects self-reported a significant increase in vocal effort following the workday. These perceptual changes coincided with an increase in vocal intensity and harmonics-to-noise ratio, but decrease in relative fundamental frequency offset 10. As expected, men and women differed in measures related to fundamental frequency and vocal tract length. Healthcare professionals wearing masks reported greater vocal symptoms post-workday compared to pre-workday. These symptoms coincided with acoustic changes previously related to vocal effort; however, the degree of change was considered mild. Further research is needed to determine whether vocal hygiene strategies may reduce vocal symptoms in mask-wearing workers.

To Nap or Not to Nap? A Systematic Review Evaluating Napping Behavior in Athletes and the Impact on Various Measures of Athletic Performance

PurposeThe objective of this systematic review was to 1) determine how studies evaluated napping behavior in athletes (frequency, duration, timing and measurement); 2) explore how napping impacted physical performance, cognitive performance, perceptual measures (eg, fatigue, muscle soreness, sleepiness and alertness), psychological state and night-time sleep in athletes.MethodsFive bibliographic databases were searched from database inception to 11 August 2020. Observational and experimental studies comprising able-bodied athletes (mean age ≥12 years), published in English, in peer-reviewed journal papers were included. The Downs and Black Quality Assessment Checklist was used for quality appraisal.ResultsThirty-seven studies were identified of moderate quality. Most studies did not include consistent information regarding nap frequency, duration, and timing. Napping may be beneficial for a range of outcomes that benefit athletes (eg, physical and cognitive performance, perceptual measures, psychological state and night-time sleep). In addition, napping presents athletes with the opportunity to supplement their night-time sleep without compromising sleep quality.ConclusionAthletes may consider napping between 20 to 90 min in duration and between 13:00 and 16:00 hours. Finally, athletes should allow 30 min to reduce sleep inertia prior to training or competition to obtain better performance outcomes. Future studies should include comprehensive recordings of nap duration and quality, and consider using sleep over a 24 hour period (daytime naps and night-time sleep period), specifically using objective methods of sleep assessment (eg, polysomnography/actigraphy).

Using ultrasound tongue imaging to analyse maximum performance tasks in children with Autism: a pilot study

ABSTRACT This study proposes a protocol for assessing speech motor control in children using maximum performance tasks with simultaneous acoustic and ultrasound recording. The protocol was piloted on eight children with autism spectrum disorders and nine typically developing children. Diadochokinesis rate, accuracy, and consistency were elicited using an imitation paradigm where speakers repeat mono-, bi-, and tri-syllabic stimuli at increasing rates. Both traditional measures of rate, accuracy and consistency and an ultrasound tongue-shape analysis of slow versus fast productions were undertaken. Preliminary results suggest that the protocol is feasible with children with communication disorders. Instrumental measures suggest greater variability in tongue movements in the children with autism that is not detected using perceptual measures of accuracy. A subgroup of children with autism showed some evidence of differences in speech motor control. Ultrasound tongue imaging appears to be a useful method for gaining additional insight into speech motor control.

A perceptual outcome measure of velopharyngeal function based on the Cleft Audit Protocol for Speech-Augmented (CAPS-A VPC-Sum): Validation through a speech osteotomy study.

The status of the velopharyngeal mechanism can be inferred from perceptual ratings of specified speech parameters. Several studies have proposed the measure of an overall velopharyngeal composite score based on these perceptual ratings and have reported good validity. The Cleft Audit Protocol for Speech-Augmented (CAPS-A) is a validated and reliable perceptual framework for the assessment of cleft speech and velopharyngeal function used by all Regional Cleft Services in the UK and Ireland. An overall velopharyngeal composite summary score based on the CAPS-A would serve as an important surgical outcome measure of speech. To develop and validate a velopharyngeal composite summary score based on perceptual ratings made on the CAPS-A (CAPS-A VPC-Sum) using data from a maxillary osteotomy (MO) study. There were two surgical groups: a cleft lip and palate (CLP) (N = 20) group and a non-CLP group (N = 10), and a normal control group (N = 20). Participants in groups 1 and 2 were seen for perceptual and instrumental assessments of speech and velopharyngeal function preoperatively (T1), 3 months (T2) and 12 months (T3) postoperatively. Perceptual speech data were collected and rated by independent listeners using CAPS-A. Moderate to strong interrater reliability for perceptual data (rs = 0.503-1.000, all p < 0.01) and strong to very strong reliability for videofluoroscopic measurements (rs = 0.746-0.947) were found. Construct validity of the CAPS-A VPC-Sum was shown by an increase in postoperative scores for the CLP group only ϰ2 (2) = 9.769, p = 0.008 and significant differences between the CLP and the other two groups at T2 and T3 using independent t-tests. Convergent and divergent validity was indicated by a positive moderate correlation with related parameters (e.g., hypernasality rs = 0.869, p < 0.01) and a weak correlation with unrelated parameters (e.g., amount of forward advancement rs = 0.160, p = 0.526). Criterion validity was found by a moderate correlation between closure ratio rs = -0.541, p = 0.020 and CAPS-A VPC-Sum. A velopharyngeal composite score based on perceptually rated parameters serves as an important surgical speech outcome measure. The CAPS-A VPC-Sum is a useful, reliable and valid outcome measure of velopharyngeal function. There are added positive implications for other clinicians using geographically and language-specific adapted versions of the CAPS-A internationally. What is already known on this subject Velopharyngeal composite scores based on perceptually rated speech parameters have been shown to have both clinical and research utility, serving as a useful surgical outcome measure. However, such a composite score must be specifically validated on the perceptual speech framework upon which it is based, as there are differences in measurement methods and terminology across cleft speech perceptual frameworks internationally. What this paper adds to existing knowledge The CAPS-A is a nationally used tool in the UK and Ireland for audit and research purposes with validated and adapted international versions. This paper reports on the validation of the derived velopharyngeal composite score measure based on the CAPS-A and an English-speaking sample, providing evidence of its validity through a speech osteotomy study. What are the potential or actual clinical implications of this work? This work provides CLP teams who use CAPS-A with a validated surgical speech outcome measure of velopharyngeal function. It has positive implications also for adapted versions of the CAPS-A internationally.

Non-traditional immersive seminar enhances learning by promoting greater physiological and psychological engagement compared to a traditional lecture format

The purpose of this study was to determine the impact of an immersive seminar, which included moderate intensities of physical activity, on learning when compared to traditional lecture format. Twenty-six healthy participants were randomly divided into an immersive seminar or traditional lecture format group and presented material related to positive psychology and human values/beliefs over the course of two days. Physical activity was collected using a bio-harness while salivary cortisol and perceptual measures of well-being were collected over the two days. Performance on an examination related to course material was used to assess learning. Average time spent over 65% of max heart rate, energy expenditure, total bounds, mechanical and physiological load were significantly greater in the immersive seminar group when compared to traditional lecture group. In addition, cortisol levels and perceptual measures of mood, focus, energy, and well-being were significantly greater in the immersive seminar when compared to the traditional lecture format. Participants in the immersive seminar demonstrated significantly greater memory retention of course material 30-days post lecture when compared to the traditional lecture group. These findings support incorporating more physical activity and increasing arousal in order to enhance learning of lecture material.

A simple vector-like law for perceptual information combination is also followed by a class of cortical multisensory bimodal neurons

SummaryAn interdisciplinary approach to sensory information combination shows a correspondence between perceptual and neural measures of nonlinear multisensory integration. In psychophysics, sensory information combinations are often characterized by the Minkowski formula, but the neural substrates of many psychophysical multisensory interactions are unknown. We show that audiovisual interactions – for both psychophysical detection threshold data and cortical bimodal neurons – obey similar vector-like Minkowski models, suggesting that cortical bimodal neurons could underlie multisensory perceptual sensitivity. An alternative Bayesian model is not a good predictor of cortical bimodal response. In contrast to cortex, audiovisual data from superior colliculus resembles the ‘City-Block’ combination rule used in perceptual similarity metrics. Previous work found a simple power law amplification rule is followed for perceptual appearance measures and by cortical subthreshold multisensory neurons. The two most studied neural cell classes in cortical multisensory interactions may provide neural substrates for two important perceptual modes: appearance-based and performance-based perception.

Effect of Combined Respiratory Muscle Training (cRMT) on Dysphonia following Single CVA: A Retrospective Pilot Study

Although dysphonia is less prevalent than dysphagia following cerebrovascular accidents, dysphonia does contribute to the burden of disease resulting from stroke. Strengthening muscles of the larynx and respiratory tract through respiratory muscle training (RMT) has proven effective in improving voice after neurological insult. However, approaches to strengthen only the expiratory muscle groups (EMST) dominate the clinical study literature, with variable outcomes. By focusing on exhalation, the contribution of inspiratory muscles to phonation may have been overlooked. This study investigated the effect of combined respiratory muscle training (cRMT) to improve voice function in stroke patients. Recorded data of twenty patients with dysphonia following stroke were allocated to an intervention (IG) or a control group (CG) based upon whether they chose cRMT or not while awaiting pro bono voice therapy services. The intervention group (n=10) was treated daily with three 5-minute sessions of combined resistive respiratory muscle training for 28 days, while the control group (n=10) received no cRMT or other exercise intervention. Perceptual and acoustic measurements as well as a pulmonary function test were assessed pre-and post-intervention. The intervention group demonstrated significant improvements after 28 days of cRMT in peak flow (127%), patient self-perception of voice improvement (84.41%), as well as in all categories of the Consensus Auditory-Perceptual Evaluation of Voice (CAPE-V): overall severity (63.22%), roughness (54.76%), breathiness (61.06%), strain (63.43%), pitch range (48.11%) and loudness (57.51%), compared to the control group who did not receive treatment. Furthermore, cRMT also led to significant improvements in maximum phonation time (212.5%), acoustic parameters of vocal intensity, and total semitone range (165.45%). This pilot study shows promise of the feasibility and effectiveness of cRMT to lessen the signs and symptoms of dysphonia while simultaneously improving breath support.

Mouth Rinsing With a Pink Non-caloric, Artificially-Sweetened Solution Improves Self-Paced Running Performance and Feelings of Pleasure in Habitually Active Individuals

Purpose: The purpose of this study was to investigate whether mouth rinsing with a pink non-caloric, artificially sweetened solution can improve self-selected running speed and distance covered during a 30 min running protocol.Methods: Ten healthy and habitually active individuals (six males, four females) completed two experimental trials in a randomised, single-blind, crossover design. Each experimental trial consisted of a 30 min treadmill run at a self-selected speed equivalent to 15 (hard/heavy) on the rating of perceived exertion scale. During exercise, participants mouth rinsed with either a pink or a clear non-caloric, artificially sweetened solution, with performance, perceptual and physiological measures obtained throughout.Results: Self-selected running speed (+0.4 ± 0.5 km·h−1, p = 0.024, g = 0.25) and distance covered (+213 ± 247 m, p = 0.023, g = 0.25) during the 30 min running protocol were both improved by 4.4 ± 5.1% when participants mouth rinsed with the pink solution when compared to the clear solution. Feelings of pleasure were also enhanced during the 30 min treadmill run when participants mouth rinsed with the pink solution, with ratings increased from 3.4 ± 0.7 in the clear condition to 3.8 ± 0.6 in the pink condition (+0.4 ± 0.5, p = 0.046, g = 0.54).Conclusion: Mouth rinsing with a pink non-caloric, artificially sweetened solution improved self-selected running speed, total distance covered, and feelings of pleasure obtained during a 30 min running protocol when compared to an isocaloric and taste-matched clear solution.

Learning digital camera pipeline for extreme low-light imaging

In low-light conditions, a conventional camera imaging pipeline produces sub-optimal images that are usually dark and noisy due to a low photon count and low signal-to-noise ratio (SNR). We present a data-driven approach that learns the desired properties of well-exposed images and reflects them in images that are captured in extremely low ambient light environments, thereby significantly improving the visual quality of these low-light images. The recent works on this problem only consider a pixel-level loss metric that ignores perceptual quality and thus generate outputs susceptible to visual artifacts. To address this problem, we propose a new loss function that exploits the characteristics of both pixel-wise and perceptual metrics, enabling our deep neural network to learn the camera processing pipeline to transform the short-exposure, low-light RAW sensor data to well-exposed sRGB images. The results show that our method outperforms the state-of-the-art according to psychophysical tests as well as pixel-wise standard metrics and recent learning-based perceptual image quality measures. In essence, the proposed model can potentially replace the conventional digital camera pipeline for the specific case of extreme low-light imaging.

Deep feature loss to denoise OCT images using deep neural networks.

.Significance: Speckle noise is an inherent limitation of optical coherence tomography (OCT) images that makes clinical interpretation challenging. The recent emergence of deep learning could offer a reliable method to reduce noise in OCT images.Aim: We sought to investigate the use of deep features (VGG) to limit the effect of blurriness and increase perceptual sharpness and to evaluate its impact on the performance of OCT image denoising (DnCNN).Approach: Fifty-one macula-centered OCT pairs were used in training of the network. Another set of 20 OCT pair was used for testing. The DnCNN model was cascaded with a VGG network that acted as a perceptual loss function instead of the traditional losses of and . The VGG network remains fixed during the training process. We focused on the individual layers of the VGG-16 network to decipher the contribution of each distinctive layer as a loss function to produce denoised OCT images that were perceptually sharp and that preserved the faint features (retinal layer boundaries) essential for interpretation. The peak signal-to-noise ratio (PSNR), edge-preserving index, and no-reference image sharpness/blurriness [perceptual sharpness index (PSI), just noticeable blur (JNB), and spectral and spatial sharpness measure (S3)] metrics were used to compare deep feature losses with the traditional losses.Results: The deep feature loss produced images with high perceptual sharpness measures at the cost of less smoothness (PSNR) in OCT images. The deep feature loss outperformed the traditional losses ( and ) for all of the evaluation metrics except for PSNR. The PSI, S3, and JNB estimates of deep feature loss performance were 0.31, 0.30, and 16.53, respectively. For L1 and L2 losses performance, the PSI, , and JNB were 0.21 and 0.21, 0.17 and 0.16, and 14.46 and 14.34, respectively.Conclusions: We demonstrate the potential of deep feature loss in denoising OCT images. Our preliminary findings suggest research directions for further investigation.