Setiawan Hadi

Benchmarking of Document Image Analysis Tasks for Palm Leaf Manuscripts from Southeast Asia

This paper presents a comprehensive test of the principal tasks in document image analysis (DIA), starting with binarization, text line segmentation, and isolated character/glyph recognition, and continuing on to word recognition and transliteration for a new and challenging collection of palm leaf manuscripts from Southeast Asia. This research presents and is performed on a complete dataset collection of Southeast Asian palm leaf manuscripts. It contains three different scripts: Khmer script from Cambodia, and Balinese script and Sundanese script from Indonesia. The binarization task is evaluated on many methods up to the latest in some binarization competitions. The seam carving method is evaluated for the text line segmentation task, compared to a recently new text line segmentation method for palm leaf manuscripts. For the isolated character/glyph recognition task, the evaluation is reported from the handcrafted feature extraction method, the neural network with unsupervised learning feature, and the Convolutional Neural Network (CNN) based method. Finally, the Recurrent Neural Network-Long Short-Term Memory (RNN-LSTM) based method is used to analyze the word recognition and transliteration task for the palm leaf manuscripts. The results from all experiments provide the latest findings and a quantitative benchmark for palm leaf manuscripts analysis for researchers in the DIA community.

Southeast Asian palm leaf manuscript images: a review of handwritten text line segmentation methods and new challenges

Abstract. Due to their specific characteristics, palm leaf manuscripts provide new challenges for text line segmentation tasks in document analysis. We investigated the performance of six text line segmentation methods by conducting comparative experimental studies for the collection of palm leaf manuscript images. The image corpus used in this study comes from the sample images of palm leaf manuscripts of three different Southeast Asian scripts: Balinese script from Bali and Sundanese script from West Java, both from Indonesia, and Khmer script from Cambodia. For the experiments, four text line segmentation methods that work on binary images are tested: the adaptive partial projection line segmentation approach, the A* path planning approach, the shredding method, and our proposed energy function for shredding method. Two other methods that can be directly applied on grayscale images are also investigated: the adaptive local connectivity map method and the seam carving-based method. The evaluation criteria and tool provided by ICDAR2013 Handwriting Segmentation Contest were used in this experiment.

Integrating computer vision technique to support Tsunami Early Warning System

Tsunami Early Warning System (TEWS) is built for minimizing impact of tsunami disaster. The generic system consists of two components: the first is sensor network for tsunami detection and the second is interconnected communication infrastructure for evacuation notification. In this paper, additional sensor based on computer vision technology is proposed for observing ocean and wave condition. This technology, combined with other sensors technology, will increase the effectiveness of the TEWS. In more detail, image frames captured from surveillance video camera are analyzed and interpreted semantically using imaging analysis techniques. If the system considered a tsunami-like situation based on image, it will transmit a warning signal to a centrally integrated system. The developed application adopted DeWa, a multiaspect framework for object detection that has been previously implemented. Although this camera-based TEWS has not been implemented in realistic situation, it has been successfully tested using extracted image from a tsunami-like wave video.

Binary similarity measure based on Moire patterns

Computing similarity measure between images is a basic task required for image classification and matching. The technique proposed explores the Moire phenomenon for determining the similarity measure for two images. When two similar images are superimposed the concentric Moire rings point to the location of the fixed point. A Hough based technique is used for circle detection. The algorithm is successfully applied to the face recognition problem.

Improved Line Segmentation Framework for Sundanese Old Manuscripts

Line segmentation can be a useful process for further text segmentation. There are some certain line segmentation framework that use binarization method as an initial step. But binarization process is still facing a major challenge, especially on old document palm-leaf manuscripts. As the quality of the image has varying degrees of noises in the non-text region. Seam Carving method, one of line segmentation methods that uses binarization-free approach, can be an alternative solution. However, this method can separate the incorrect text line on small element text located at the bottom or at the top of a main character contour. Therefore, an improvement on line segmentation framework is proposed by using hybrid binarization and its implemented on the smallest energy function to separate out the text-lines. The proposed framework have been evaluated on 44 Sundanese old manuscript images that consist of true color and binary images. The evaluation matrix shows that this framework can improve Niblack binarization process up to 50%. In addition, our framework does not only generate the number of text-lines to come near to the number of target lines, but it also can separate the text-lines well on small element text. Overall, the expected result can in the end be produced from the proposed line segmentation framework.

The design and preliminary implementation of low-cost brain-computer interface for enable moving of rolling robot

We introduce the design and preliminary implementation of low-cost brain-computer interface (BCI) to enable movement of a rolling robot. This system will accept and execute basic commands that are generated from three brain condition normal, relax, or happy. The brain condition is determined by brainwaves known as alpha, beta, and gamma waves. The brainwaves are detected using Mindflex, connected with Arduino Uno, and sent data transmission to computer via USB connection. Algorithm is developed using Matlab to analyze the data and send three simple commands to a rolling robot through Wi-Fi connection. We conducted experiment for fifty times using 100 data for training and 50 data for testing, and obtained 62% of accuracy. This result shows that brainwave commands can be processed successfully for forward, turn left, and turn right movements. We have concluded that the system can be developed further to assist disabilities performing motions using their minds.

Sundanese ancient manuscript retrieval system comparison of two probability approaches

Nowadays, the development of information technology both in terms of hardware and software has been growing rapidly. Many people used technology to search many types of information. On the other hand, there are information on Sundanese ancient manuscript as cultural heritage with many valuable information still accessed manually, such as the manuscript which located in Situs Kabuyutan Ciburuy, Garut, Jawa Barat. Many people with many purposes (i.e. Research, education, etc.) came to access directly. Its becoming a problem because due to lack of maintainability, the manuscripts became fragile. Furthermore, the information will be lost. So, the emergency solution should be proposed to preserve this valuable heritage. We proposed a retrieval system using the probability approach that can be effective for search the information in ancient Sundanese language. The two probability approaches: Bayesian Networks and Divergence form Randomness have been compared. The 50 queries for three different types of structured query have been run and evaluated by Mean Average Precision (MAP). The results give that the effective methods come from Bayesian Network. There is another finding during the building process of retrieval system. The interactive visualization can be delivered by the retrieval system according to the given query.

Converting images to pseudo random dots: performance evaluation

Applications of Moire patterns have been reported in stereo matching (disparity measurement) and in developing a binary similarity measure. Matching based on Moire patterns requires that the images be converted to pseudo random dot representation. Two classes of algorithms for converting images to random dots are presented: (i) high pass filtering, and (ii) bit slicing. Parameters which influence dot density and uniformity of distribution are investigated.