![]() These annotations will be worse than those that would be obtained using a commercial OCR system (like those used to pre-train LayoutLMv2/v3). These are the annotations used to pre-train Dessurt ( ). Important: There is also a "rotation" value in the json (0, 90, 180, or 270) indicating the json may be for a rotated version of the IIT-CDIP image by the given amount (attempted to rotated documents to upright position to get better OCR results). the problem is that, thats all i get regarding the error: E/AbstractTracker: Cant create handler inside thread that has not called Looper.prepare() D/AppTracker: App Event: stop E/AbstractTracker: Cant create handler inside thread that has not called Looper. The combine output of these models became the block/paragraph annotations (we kept the Tesseract output format, but each block has 1 paragraph of exactly the same shape). The images were then run through both the Publaynet and PrimaNet models available on LayoutParser ( ). The block and paragraph output of Tesseract was discarded. The line and word annotations are directly taken from Tesseract. The jsons contain block/paragraph, line and word bounding boxes, with transcriptions for the words following the Tesseract format. This dataset contains a "X.layout.json" for each "X.png" in the IIT-CDIP dataset (doesn't have sections 'a', 'w', 'x', 'y', and 'z'). directories) and can thus be combine with the image IIT-CDIP dataset using rsync or similar tool. The directory struture of this dataset is the same as the IIT-CDIP dataset (although has everything in one tar, with "a.a", "a.b". To download the images of the IIT-CDIP dataset go to 10 of 12 resondents to a post-trial questionnaire felt the app helped them manage their smartphone usage to at least a decent extent.This is Tesseract generated transcriptions (no images) of (most of) the IIT-CDIP dataset. Additionally, average daily usage of apps the user did not choose to use less reduced by only 1.7%. The results were very pleasing: analysis found that average daily usage of apps each user chose to use less reduced by 25.1% on average for the period after installing AppTracker, which was a statistically significant result. Anonymised usage data was collected using Firebase. More on the design of the app and the theory behind it can be found in the project report.Īfter the app was developed, a two week study was completed by fourteen participants who used the app on their phones. Secondly, the glow scales to the amount of usage, being barely noticeable after a short time, but gradually becoming more and more obtrusive. Useful functions of the phone remain fully accessible. Firstly, it is targeted in that only apps the user has said they want to use less of are affected - so the user has already bought into this. This is a targeted and scaled approach, aimed at reducing the frustration experienced by the user as a result of using AppTracker, as frustration is the main cause of deactivating similar intervention software. The app therefore asks users to select apps they wish to use less of, and to set a daily usage target. In short, the edge of the screen gets more and more red when a problematic app is used for a longer period of time, and the more these apps are used in a single day, the more of the screen the glow will surround. ![]() The latter builds on a finding from previous research that these long, mindless, habitual use cases are perceived as particularly unsatisfying by users, giving a sense of a lack of autonomy. In the improved P4P system, a distributed tracker overlay network replaces the appTracker to manage the resources in the different ISP domains. ![]() It changes into two dimensions - the amount of the screen edge it covers, and the intensity/colour/brightness of the glow - to communicate two distinct quantities the user's progress towards a daily target of usage, and particularly intense usage sessions. An improved design of P4P based on distributed tracker is proposed to solve the overload problem of single appTracker server. This awareness is achieved by displaying an 'edge glow' on screen, at all times. AppTracker attempts to remove the need for frustrating interventions such as outright restriction of app usage, by instead giving the user a continuous, subliminal awareness of their smartphone usage, encouraging mindfulness. A survey of existing literature and solutions found that while tracking of phone use is now a common and readily available feature, this tracking data is not always immediately available to the user. Tackling problematic smartphone usage with on-screen usage feedback Or 'helping people use their phones less by giving them a sense of how much they've used their device'ĪppTracker is an app developed as part of an MSc Human Computer Interaction dissertation, which builds on research into problematic smartphone usage and techniques.
0 Comments
Leave a Reply. |