This will count as one of your downloads.
You will have access to both the presentation and article (if available).
Current developments in sensors open new possible uses across numerous real-life applications, including optical character recognition (OCR). An OCR system requires incorporation of text processing tools into the sensor functionality. The most critical stage in OCR systems is the segmentation stage. It refers to the challenge of subdividing a text image into characters, which can be individually processed using a classifier. The cursive nature of the Arabic script such as the existence of different shapes for each character according to its location in the word besides the existence of diacritics makes Arabic character segmentation a very challenging task. A robust offline character segmentation algorithm for printed Arabic text with diacritics is developed based on the contour extraction technique. The algorithm works through extracting the up-contour part of a word and then identifies the splitting areas of the word characters. Then a postprocessing stage is used to handle the over-segmentation problems that appear in the initial segmentation stage. The proposed scheme is benchmarked using the APTI dataset and a manually collected dataset consisting of image texts varying in font size, type, and style for more than 38,000 words. The experiments show that the proposed algorithm is able to segment Arabic words with diacritics with an average accuracy of 98.5%.
This will count as one of your downloads.
You will have access to both the presentation and article (if available).
Accurate environmental perception is essential for ensuring the safe operation of autonomous vehicles and robots. However, unpredictable factors like illumination, noise, and adverse weather conditions can significantly impact the performance of autonomous technologies, particularly in critical applications such as navigation and self-driving. Deep learning is one potential solution for object detection and scene
perception problems, which can enable algorithm-driven and data-driven cars. This course introduces Quaternion Neural Networks (QNNs), a novel framework for robust scene perception, especially in adverse weather conditions. QNNs leverage quaternion numbers to process multidimensional inputs, capture internal dependencies, and improve scene segmentation and object detection under challenging scenarios. The talk covers the basics, such as quaternion convolution and QNN transformers. It highlights their application in scene perception for self-driving cars under foggy, cloudy, and rainy environments. Attendees will see how QNNs surpass real-valued networks in weather removal, scene segmentation, and detection tasks. It will also discuss the recent trends in these technologies and the associated commercial impact and opportunities.
Artificial Intelligence (AI) has become everywhere in our society, with applications in various fields such as search engines, image recognition, security systems, medical diagnosis, drones, and self-driving cars. One of the core areas of AI is computer vision, which involves teaching machines to interpret and understand images and videos. This course will introduce cutting edge machine learning and deep learning techniques, focusing on creating and customizing convolutional neural networks (CNNs) in real and quaternion domains. The course will cover (i) the fundamental components that drive modern deep learning systems for computer vision, (ii) the possibilities and limitations of deep learning methods, and (iii) the applications of deep learning to various problems in computer vision, including emerging topics.
Processing, transmission, storage, and distribution of digital multimedia data have become the major tasks of today's Internet and operating systems, and will continue to be the major driving strength to networking and systems research in the future. Modern advancements in information technology have enabled pervasive uses of digital multimedia data in a variety of scientific, government, business, and consumer applications. The digital nature of the information also allows individuals to manipulate, duplicate or access information beyond the terms and conditions agreed upon in a given transaction. Multimedia security involves the protection of digital audio, video and images against illegal access, tampering, piracy, storage, and transmission.
The goal of the course is to introduce the theory and methods of multimedia communication & security systems: these include encryption, steganographic (information hiding), and watermarking systems. The following major areas will be covered in this class – Multimedia representation systems and file formats; state-of-the-art methods and techniques in this area; multimedia security systems applications. Future directions and open problems/issues will be also discussed.
View contact details
No SPIE Account? Create one
