We provide normative data for a new set of 313 colourised line drawings. The drawings were standardised on name agreement ( N = 60 participants), image agreement ( N = 34), conceptual familiarity ( N = 36), age of acquisition ( N = 35), and imageability ( N = 35). Objective visual complexity measures are given for the pictures, and objective word frequencies are provided for the modal names of the drawings. Reliability measures for the collected norms are very high. There are high levels of agreement between the names given by the participants and the drawings and comparative analyses indicate that the distribution of name agreement scores is very similar in both our own database and the MultiPic database (Duñabeitia et al., 2018). A novel “picture-choice task” used to assess name-image agreement ( N = 30) reveals that the great majority of the IMABASE pictures that are also present in MultiPic are rated as providing better pictorial representations of the corresponding concepts. Finally, most of the correlations are comparable with those reported in other normative studies on colourised drawings. The whole set of pictures is freely available from https://leadserv.u-bourgogne.fr/~lead/imabase/ and the norms are available as Supplementary Material.
Text line segmentation is a necessary preliminary step before most text transcription algorithms are applied. The leading deep learning networks used in this context (ARU-Net, dhSegment, and Doc-UFCN) are based on the U-Net architecture. They are efficient, but fall under the same concept, requiring a post-processing step to perform instance (e.g., text line) segmentation. In the present work, we test the advantages of Mask-RCNN, which is designed to perform instance segmentation directly. This work is the first to directly compare Mask-RCNN- and U-Net-based networks on text segmentation of historical documents, showing the superiority of the former over the latter. Three studies were conducted, one comparing these networks on different historical databases, another comparing Mask-RCNN with Doc-UFCN on a private historical database, and a third comparing the handwritten text recognition (HTR) performance of the tested networks. The results showed that Mask-RCNN outperformed ARU-Net, dhSegment, and Doc-UFCN using relevant line segmentation metrics, that performance evaluation should not focus on the raw masks generated by the networks, that a light mask processing is an efficient and simple solution to improve evaluation, and that Mask-RCNN leads to better HTR performance.
27 August 2022 in International Journal of Child-Computer Interaction
Mental fatigue is common in society, but its effects on force production capacities remain unclear. This study aimed to investigate the impact of mental fatigue on maximal force production, rate of force development-scaling factor (RFD-SF), and force steadiness during handgrip contractions. Fourteen participants performed two randomized sessions, during which they either carried out a cognitively demanding task (i.e., a visual attention task) or a cognitively nondemanding task (i.e., documentary watching for 62 min). The mental fatigue was evaluated subjectively and objectively (performances and electroencephalography). Maximal voluntary contraction (MVC) force, RFD-SF, and force steadiness (i.e., force coefficient of variation at submaximal intensities; 25, 50, and 75% of MVC) were recorded before and after both tasks. The feeling of mental fatigue was much higher after completing the cognitively demanding task than after documentary watching (p < .001). During the cognitively demanding task, mental fatigue was evidenced by increased errors, missed trials, and decreased N100 amplitude over time. While no effect was reported on force steadiness, both tasks induced a decrease in MVC (p = .040), a force RFD-SF lower slope (p = .011), and a reduction in the coefficient of determination (p = .011). Nevertheless, these effects were not explicitly linked to mental fatigue since they appeared both after the mentally fatiguing task and after watching the documentary. The study highlights the importance of considering cognitive engagement and mental load when optimizing motor performance to mitigate adverse effects and improve force production capacities.
1 March 2025 in Journal of Experimental Psychology: Learning, Memory, and Cognition
Classical music pianists of five different conservatory levels, from undergraduate to professional, were tested on a sight-reading task with eye-movement recording. They had to sight read both tonal classical scores that followed the rules specific to Western tonal music, and atonal contemporary scores, which do not follow these rules. This study aimed at determining the extent to which eye movements and musical performance metrics can account for the level of sight-reading expertise. First, the results indicated that with the acquisition of expertise, musicians process visual information more rapidly (increasing their played tempo while decreasing average fixation duration and their number of fixations), more structurally (tending to increase their eye-hand span), and more accurately (increasing their sight-reading accuracy). Second, when they sight read contemporary scores compared to classical scores, musicians decreased their played tempo, tended to be less accurate, increased their number of fixations, and tended to decrease their eye-hand span. Finally, expertise effects were moderated by the type of score. These results suggest (a) that visual perception is progressively shaped through music reading expertise and through domain-specific knowledge acquisition, (b) that tonal-specific cues play a significant role to use an efficient eye-movement behavior and (c) that the benefit conferred by expert prior music-specific knowledge seems to be even greater for sight-reading tonal rather than atonal scores. Our findings are discussed in the light of expert memory theories (long-term working memory theory; Ericsson & Kintsch, 1995; template theory, Gobet & Simon, 1996). (PsycInfo Database Record (c) 2024 APA, all rights reserved).