DeHyFoNet: Deformable Hybrid Network for Formula Detection in Scanned Document Images

This work presents an approach for detecting mathematical formulas in scanned document images. The proposed approach is end-to-end trainable. Since many OCR engines cannot reliably work with the formulas, it is essential to isolate them to obtain the clean text for information extraction from the document. Our proposed pipeline comprises a hybrid task cascade network with deformable convolutions and a Resnext101 backbone. Both of these modifications help in better detection. We evaluate the proposed approaches on the ICDAR-2017 POD and Marmot datasets and achieve an overall accuracy of 96% for the ICDAR-2017 POD dataset. We achieve an overall reduction of error of 13%. Furthermore, the results on Marmot datasets are improved for the isolated and embedded formulas. We achieved an accuracy of 98.78% for the isolated formula and 90.21% overall accuracy for embedded formulas. Consequently, it results in an error reduction rate of 43% for isolated and 17.9% for embedded formulas.

Attacks to Automatous Vehicles: A Deep Learning Algorithm for Cybersecurity

Rapid technological development has changed drastically the automotive industry. Network communication has improved, helping the vehicles transition from completely machine- to software-controlled technologies. The autonomous vehicle network is controlled by the controller area network (CAN) bus protocol. Nevertheless, the autonomous vehicle network still has issues and weaknesses concerning cybersecurity due to the complexity of data and traffic behaviors that benefit the unauthorized intrusion to a CAN bus and several types of attacks. Therefore, developing systems to rapidly detect message attacks in CAN is one of the biggest challenges. This study presents a high-performance system with an artificial intelligence approach that protects the vehicle network from cyber threats. The system secures the autonomous vehicle from intrusions by using deep learning approaches. The proposed security system was verified by using a real automatic vehicle network dataset, including spoofing, flood, replaying attacks, and benign packets. Preprocessing was applied to convert the categorical data into numerical. This dataset was processed by using the convolution neural network (CNN) and a hybrid network combining CNN and long short-term memory (CNN-LSTM) models to identify attack messages. The results revealed that the model achieved high performance, as evaluated by the metrics of precision, recall, F1 score, and accuracy. The proposed system achieved high accuracy (97.30%). Along with the empirical demonstration, the proposed system enhanced the detection and classification accuracy compared with the existing systems and was proven to have superior performance for real-time CAN bus security.

Kinerja Jaringan Sensor Nirkabel untuk Model Smart Building

Penggunaan energi pada jaringan sensor nirkabel saat ini bisa dikatakan sangat boros, sehingga dibutuhkan sebuah metode komputasi pada teknologi Internet of Things (IoT) dengan sumber yang terbatas. Konsep penggunaan energi pada IoT perlu diawasi dan dikelola supaya terdapat peningkatan efisensi penggunaan energi sehingga dapat menekan biaya tanpa harus mengurangi kinerjanya. Dalam penelitian ini dibuatlah sebuah prototype sistem kontrol penggunaan energi pada IoT yang nantinya digunakan untuk model Smart Building dalam upaya penghematan energi. Konsep prototipe yang dibuat adalah hybrid network. Pada prototipe digunakan beberapa sensor berfungsi untuk membaca, mengontrol, dan mengirimkan informasi secara realtime dengan protokol zigbee IEEE 802.15.4 dan Wi-Fi ESP8266-01. Model ini diamati dan dianalisis konsumsi energi dan Quality of Service (QoS) transmisi pengiriman data dengan metode clustering. Kinerja pada jaringan bekerja dengan baik dilihat dari kondisi pada lingkungan indoor maupun outdoor. Pengujian konsumsi energi seluruh proses pengiriman data sensor yaitu 0.32 Watt sampai dengan 0.64 Watt dalam waktu 10 Menit, dapat dikatakan komunikasi jaringan pada pengembangan hybrid network ini mengonsumsi energi yang sangat rendah. Total durasi waktu pengujian dengan baterai 12 Volt hingga batas tegangan minimum didapatkan 70 Menit dan tegangan minimal mencapai 7 Volt.

Cultural Magnitude: Interactive visual mapping - A hybrid network for indigenous data

<p>This research has been undertaken in response to the limitations of standard mapping techniques, in particular, those that use ESRI-based technology and delivery. The work argues that our ability to understand the complex nature of indigenous ontologies and spatial models are affected by the available tools and their ontological frameworks. It sets out to visualise, in a tool, traditionally non-physical, but inherently spatial, data and information. The map, in a traditional sense, now becomes a fluid, open, self-referential virtual topography or ‘space’, challenging the rational top-down fixity of western cartographic representation. As an architectural thesis, it seeks to create holistically structured space as a virtual edifice and is concerned with that which is not represented and concludes that the most important aspect of creating a mapping framework for an indigenous ontology is to understand the inseparable relationship between people, knowledge and land. The research describes a tool designed and built by the author that contributes to cultural and spiritual health (whai ora) and wellbeing of Māori. Through its ontological framework, it aims to provide an alternate map that enables users to navigate and share cultural knowledge. The central concept is to ‘re-connect’, in particular, urban and disenfranchised Māori, through the creation of a virtual space that can be customised and inhabited in various ways by its users. It questions and challenges what is included and what is excluded, what can be represented, asking where might culture have a ‘place’? How might people and their environments effect change in themselves? In others? Cultural Magnitude is the exploration of the development of a tool that acts as a digital representation and storage place of whakapapa and taonga, and as a cultural resource for Māori to understand their spiritual bounds to physical locations - a tangible foundation for a digital marae.</p>