GroupGazer: A Tool to Compute the Gaze per Participant in Groups with integrated Calibration to Map the Gaze Online to a Screen or Beamer Projection

Fuhl, Wolfgang

doi:10.48550/arxiv.2201.07692

Cited by 8 publications

(13 citation statements)

References 78 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…DNNs [30] have found their way into a variety of fields [1,27]. In eye tracking [31,49], they are already used for scanpath analysis [32,33,4], as well as other approaches based on machine learning [24,2,3,12], feature extraction [49] such as pupil [28,18,25,48,47,46,11,26,16,10,8,45,6], iris [9,19,5] and eyelids [22,21,23], eyeball estimation [7], and also for eye movement classification [41,42,14]. In recent years, there have been a number of new large data sets [38] that have also been annotated using modern machine learning approaches.…”

Section: Introductionmentioning

confidence: 99%

Technical Report: Combining knowledge from Transfer Learning during training and Wide Resnets

Fuhl¹

2022

Preprint

Self Cite

View full text Add to dashboard Cite

In this report, we combine the idea of Wide ResNets and transfer learning to optimize the architecture of deep neural networks. The first improvement of the architecture is the use of all layers as information source for the last layer. This idea comes from transfer learning, which uses networks pre-trained on other data and extracts different levels of the network as input for the new task. The second improvement is the use of deeper layers instead of deeper sequences of blocks. This idea comes from Wide ResNets. Using both optimizations, both high data augmentation and standard data augmentation can produce better results for different models. Link:

show abstract

Section: Introductionmentioning

confidence: 99%

Technical Report: Combining knowledge from Transfer Learning during training and Wide Resnets

Fuhl¹

2022

Preprint

Self Cite

View full text Add to dashboard Cite

show abstract

“…Preto bohužia ľ nemôžeme predpoklada ť, že Λ ι nie je odlišná od p. Chceme rozšíri ť výsledky [76,10] na degenerované kruhy. Tu je jednoznačne dôležitá jednoznačnos ť. Už dlho je známe, že j ∼ E [76,45,43,39,52]. V článku [1] sa autori zaoberajú spojitos ťou homomorfizmov za ďalšieho predpokladu, že √ 2 ≤ 12.…”

unclassified

Konvexné skaláre pre uzavretý pravouhlý systém

Arsenio¹

2022

Preprint

View full text Add to dashboard Cite

show abstract

“…stimulation [38]. A number of studies have linked fixation-related metrics to cognitive effort [23,34,44]. For example, the number of fixations within a study area (AOI) was used for comparison.…”

mentioning

confidence: 99%

“…In order to pay attention to a stimulus or object, the user must expend effort to maintain maintain a constant gaze on the object [20]. Furthermore, studies provide evidence that fixation duration increases as the information processing becomes more demanding [34,57,78].…”

mentioning

confidence: 99%

See 1 more Smart Citation

Task Demand evaluation with Machine Learning in eye tracking

Cummings¹

2022

Preprint

View full text Add to dashboard Cite

Information systems play an important role in today’s rapidly evolving digital economy. As task demands are amajor factor influencing how IT systems are used for decision making, it is necessary to identify the tasks that areimportant to them. Requirements automatically provide an opportunity to design advanced decision support systemsthat can respond to the needs of users at a customised level. A first step in designing such advanced decision supporttools is, for example Exploring the possibility of developing automatic task detectors. Based on the decision makingprocess, the eye eye tracking and machine learning literature, we argue that task difficulty can be reliably detectedautomatically, and unobtrusively, using only eye movements. To explore this possibility, we developed an eye trackingsystem. Task difficulty detection system and tested its effectiveness. Our results showed that our task task task detectionsystem reliably predicted increased task difficulty based on user eye movement data. These results and for research andpractice.Recent advances in specialized hardware and software provide the ability to capture and analyze physiologicalmeasurements that can reliably detect and potentially predict changes in user behaviour. This opportunity, in turn,makes it possible to imagine intelligent decision support systems that can more more effectively recognize and respondto user needs at a personalized level [6,30]. The first step towards developing such adaptive decision support systems isto explore intelligent tools that can be reliable and unobtrusive detect user needs in a reliable and unobtrusive way. Onesuch tool is a system that can automatically detect task demand. Task demand plays a significant role in how peoplechoose to process information and therefore has a significant impact on how individuals use computers to make decisions[62, 74-76]. Because computers can extend an individual’s capabilities to process information, it is often assumed thatthey can help people improve their decision-making. However, studies report that this is not always the case. Studiesshow that people often use computers in a way to reduce their effort rather than maximize their accuracy [74-76]. Suchtechnology use behaviour is not due to innate laziness or indifference; while precise and rational decisions represent theintention, by limited cognitive capacity.

show abstract

GroupGazer: A Tool to Compute the Gaze per Participant in Groups with integrated Calibration to Map the Gaze Online to a Screen or Beamer Projection

Cited by 8 publications

References 78 publications

Technical Report: Combining knowledge from Transfer Learning during training and Wide Resnets

Technical Report: Combining knowledge from Transfer Learning during training and Wide Resnets

Konvexné skaláre pre uzavretý pravouhlý systém

Task Demand evaluation with Machine Learning in eye tracking

Contact Info

Product

Resources

About