Research Interests : Remote Sensing Image Recognition, And Machine Learning

Some more advanced methods combine Top-down and bottom-up philosophy. More specificcaly [15] Et al proposed a new text line location and separation algorithm for complex handwritten documents that is based on the application of a fuzzy directional runlength. Intuitively, the fuzzy runlength at a pixel is how far we can see from standing at the pixel along horizontal direction, by skipping some foreground pixels until a threshold (20 to 25 pixels). A horizontal fuzzy runlength map is generated by scanning every row from left to right, and right to left and keeping the fuzzy runlegth at each pixel. The fuzzy runlength map is then binarized to expose the location of text items that are part of text lines. Finally using a heuristic approach text items are grouped into text lines based on their

I’m Jon Genty, at the University of Houston, where we’re interested in the way inventive minds work.

• I oversee and assist over 100 investigators conducting more than 250 projects within the Research Department.

Although I am still just an undergraduate, I have had the great fortune to already participate in the academic community through research. Specifically, I have been worked for my P.I., Tim P. Szczykutowicz, within the medical physics department of the Wisconsin Institute

I have conducted research at two different research groups. One of the research groups was the BioMechanics Laboratory at NYU Abu Dhabi. During my sophomore summer, I developed a proof of concept prototype used to measure the flexural rigidity of cornstalk at that group. Professor Douglas Cook was my faculty mentor while Daniel Robertson was my post-doc supervisor during that time. The other research group where I am still working at is the Dynamical Systems Laboratory at NYU Tandon School of Engineering. I started conducting research at the DSL laboratory during my junior fall. The purpose of my research at the DSL group is to develop a visual based autonomous navigation system for an Aquatic Robotic Vehicle that is used to monitor the water quality of the Brooklyn Gowanus Canal in the Brooklyn Atlantis citizen science project. Professor Porfiri is my faculty mentor while Jeffrey Laut was my PhD supervisor at this group.

For more than thirty years, reaserchers have been working on handwritten recognition. Over the past few years, the number of companies involved in research on handwritten recognition has continually increased. The advance of handwritten processing results from a combination of various elements,

The outline of the stages of the image processing pipeline is; realign (affine), skull scripting, statistical edge detector (multi-scale, multi-modal), extract statistical features, use it to build SVM. The robustness of implementation will be tested with Pixel Correspondence Metric (PCM). The same pipeline will be followed for an unsupervised approach using leave one out (LOOCV). The filter size, parametric and non-parametric features like mean, variance, rank, U Mann Whitney test etc will be evaluated and an optimal performer will be selected.

Q: Another field which you showed interest in was optics. What kind of research did you undertake, and did you ever get a work in this study published?

I currently work for Dr. Garcia-Garibay, designing a new diverging synthetic pathway in order to create rigid, highly symmetrical molecules called pentiptycene. Optimizing this strategy will allow easier access to these molecules that can be hopefully used for molecular machines, supramolecular chemistry, and polymers. Thanks to this research experience, I have join UC LEAD as a scholar. Thanks to this program, I have presented in five poster presentations including The Annual Biomedical Research Conference for Minority Students (ABRCMS) in Tampa, Florida. Another component to this program is doing research at a different

My professional background is in business management and entrepreneurship. The four years I spent in Hawaii 's biotechnology industry improved my ability to think analytically and provided me with experience in applied scientific research. While working in biotechnology research, I regularly collaborated with Ph.D. level colleagues to produce professional quality results under strict deadlines.

On Amir On joined MIT as a PhD student a year after me and became “my” first student. As my first student, On had a tremendous role in shaping what I expect from students and how I see the professor-student relationship. In addition to being exceptionally smart, On has an amazing set of skills, and what he does not know he is able to learn within a day or two. It is always exciting to work and spend time with him. On is currently a professor at the University of California at San Diego. Marco Bertini When I first met Marco, he was a PhD student at Harvard Business School, and unlike his fellow students he did not see the Charles River as an obstacle he should not cross. Marco is Italian, with a temperament and sense of style to match—an overall great guy you just want to go out for a drink with. Marco is currently a professor at London Business School. Ziv Carmon Ziv was one of the main reasons I joined Duke’s PhD program, and the years we spent together at Duke justified this decision. Not only did I learn from him a great deal about decision making and how to conduct

The use of remote sensing to observe weather patterns especially hurricanes and their aftermath have only been widely used in the last fifty years. Early hurricane observations were made from ships or on land up until the 1960s (Hodgson, Davis, Cheng, Miller, 2013, p. 9). During that decade, the first United States satellite capable of producing images, the Television Infrared Observation Satellite (TIROS), was launched into space (Hodgson et. al, 2013, p. 9). This allowed for meteorologists to observe the movement and strength of tropical systems that were difficult to track without overhead imagery. Since then the technology and widespread use of these satellites have only increased and have become a major part of everyday life for people. A hurricane can create immense destruction from its strong winds, storm surge, and rain. They also can cause large losses of life which render being able to track and prepare for these systems even more vital. Remote sensing has become essential in protecting the safety of civilians from natural disasters. Satellites can help organizations like the National Oceanic and Atmospheric Administration (NOAA) track these storms, create models for potential strengthening and their paths, and inform the population so local governments, state governments, and the individual can begin to prepare. Additionally, the importance of disaster relief in the aftermath of a hurricane is great given the amount people who are in danger and need medical

Abstract : The Objective of this project is to recognise english handwriting from a given document image. In this report we have used 40point feature extraction to extract pattern from the characters and then use this data to train artificial neural network. This system is very successful in recognition of handwritten characters. Hence this system will be suitable to convert handwritten text into text document.

Segmentation of handwritten document images into text-lines and words is a necessary job for optical character recognition. Conversely, since the topographies of handwritten document are asymmetrical and miscellaneous reliant on the person, it is considered a thought-provoking problem. To address the problem, we articulate the word segmentation problem as a Binary Quadratic Assignment (BQA) problem that contemplates pairwise correspondences amongst the gaps in addition to the likelihoods of distinct gaps. Although many attributes are involved in our articulation, we guesstimate all attributes depending on the Structured SVM (Support vector machine) framework so that the anticipated method works fine irrespective of writing styles and written languages deprived of user-0 defined attributes. Experimental results on ICDAR 2009/2013 handwriting segmentation databases displays that anticipated method attains the hi-tech performance on

Most of the Optical Character Recognition (OCR) system works on a critical assumption on the script used in the image document that is supplied for recognition. A falsely selected choice of language or script type will hinder the performance of the OCR system. Therefore human intervention is required to select the appropriate package related to the supplied documents. This approach is certainly inappropriate, inefficient, impractical and undesirable. An intermediate process of script identification is required to be appended after the normal preprocessing step of skew correction, resizing, cropping and binarization. The output of this script identification process helps to determine the script used in the documents, and thus human intervention can be eliminated. Automatic script identification will not only enable to identify the script, but it can be further implemented for archiving work such as sorting and searching of document image for a