Deep learning has been proven be very effective for various image recognition tasks, e.g., image classification, semantic segmentation, image retrieval, shape classification etc. However, existing works on deep learning for image recognition mainly focus on either natural image data or binary shape data. In this paper, we show that deep convolutional neural networks (DCNN) is also suitable for ...

We present efficient streaming algorithms to compute two commonly used anomaly measures: the rank-k leverage scores (aka Mahalanobis distance) and the rank-k projection distance, in the rowstreaming model. We show that commonly used matrix sketching techniques such as the Frequent Directions sketch and random projections can be used to approximate these measures. Our main technical contribution...

We propose and test several computational methods to automatically determine possible saliency cut-off points in Sketch Engine (Kilgarriff and Tugwell, 2001). Sketch Engine currently displays collocations in descending importance, as well as according to grammatical relations. However, Sketch Engine does not provide suggestions for a cut-off point such that any items above this cut-off point ma...

The paper describes work on writing a Russian Sketch grammar for the system Sketch Engine. The objective of such a system is to provide lexicographers with sufficient lexical material and tools for getting information about a word’s collocability and to generate lists of the most frequent phrases for a given word, and then to classify them for appropriate syntactic models. The system will give ...

This paper shows the potential of using a shape outline matching technique that is based on the Turning Function distance measure in the process of creating 2D graphics. The user of a sketch-designing application is able to utilize shape-matching technology to increase the ease and correctness of the multimedia product, which is the resulting sketch or image. The shape-matching method described...

This research describes efforts to expand the lexical resource VerbNet with additional class members and completely new verb classes. Several approaches to this in the past have involved automatic methods for expansion, but this research focuses on the addition of frequent, yet particularly challenging verbs that require manual additions after a survey of each verb’s syntactic behaviors and sem...

Blake Williford Sketch Recognition Lab Texas A&M University [email protected] Wayne Li School of Industrial Design Georgia Institute of Technology [email protected] Ethan Hilton iDREEM Lab Georgia Institute of Technology [email protected] Julie Linsey iDREEM Lab Georgia Institute of Technology [email protected] Tracy Hammond Sketch Recognition Lab Texas A&M University tham...

Computer Sketch Recognition

What: Sketch recognition is a natural form of interaction that has been receiving increasing attention in the research community [4], [1], [6]. One of the challenges in sketch recognition is efficient recognition. We propose to create efficient recognition by treating sketching as a dynamic process and by deriving recognition strategies from the object models specified in the language proposed ...

iCanDraw is a drawing tool that can assist novice users to draw. The goal behind the system is to enable the users to perceive objects beyond what they know and improve their spatial cognitive skills. One of the early tasks in a beginner art class is to accurately reproduce an image, in an attempt to teach users to draw what they see, rather then what they know, improving spatial cognition skil...