Hand gesture recognition serves as a key for overcoming many difficulties and providing convenience for human life. The ability of machines to understand human activities and their meaning can be utilized in a vast array of applications. One specific field of interest is sign language recognition. This paper provides a thorough review of state-of-the-art techniques used in recent hand gesture and sign language recognition research. The techniques reviewed are suitably categorized into different stages: data acquisition, pre-processing, segmentation, feature extraction and classification, where the various algorithms at each stage are elaborated and their merits compared. Further, we also discuss the challenges and limitations faced by gesture recognition research in general, as well as those exclusive to sign language recognition. Overall, it is hoped that the study may provide readers with a comprehensive introduction into the field of automated gesture and sign language recognition, and further facilitate future research efforts in this area.

A review of hand gesture and sign language recognition techniques

We present a new approach to rapid prototyping of functional objects, such as the body of a head-mounted display. The key idea is to save 3D printing time by automatically substituting sub-volumes with standard building blocks -- in our case Lego bricks. When making the body for a head-mounted display, for example, getting the optical path right is paramount. Users thus mark the lens mounts as "high-resolution" to indicate that these should be 3D printed. faBrickator then 3D-prints only these parts. It also generates instructions that show users how to create everything else from Lego bricks. If users iterate on the design later, fa-Brickator offers even greater benefit as it allows re-printing only the elements that changed. We validated our system at the example of three 3D models of functional objects. On average, our system fabricates objects 2.44 times faster than traditional 3d printing while requiring only 14 minutes of manual assembly.

faBrickation: fast 3D printing of functional objects by integrating construction kit building blocks

In this paper, different well-known features for image retrieval are quantitatively compared and their correlation is analyzed. We compare the features for two different image retrieval tasks (color photographs and medical radiographs) and a clear difference in performance is observed, which can be used as a basis for an appropriate choice of features. In the past a systematic analysis of image retrieval systems or features was often difficult because different studies usually used different data sets and no common performance measures were established.

Features for image retrieval: A quantitative comparison

A majority of deaf 18-year-olds in the United States have an English reading level below that of a typical 10-year-old student, and so machine translation (MT) software that could translate English text into American Sign Language (ASL) animations could significantly improve these individuals' access to information, communication, and services. Previous English-to-ASL MT projects have made limited progress by restricting their output to subsets of ASL phenomena---thus avoiding important linguistic and animation issues. None of these systems have shown how to generate classifier predicates (CPs), a phenomenon in which signers use special hand movements to indicate the location and movement of invisible objects (representing entities under discussion) in space around their bodies. CPs are frequent in ASL and are necessary for conveying many concepts. 
This project has created an English-to-ASL MT design capable of producing classifier predicates. The classifier predicate generator inside this design has a planning-based architecture that uses a 3D "visualization" model of the arrangement of objects in a scene discussed by the English input text. This generator would be one pathway in a multi-path English-to-ASL MT design; a separate processing pathway would be used to generate classifier predicates, to generate other ASL sentences, and to generate animations of Signed English (if the system lacked lexical resources for some input). 
Instead of representing the ASL animation as a string (of individual signs to perform), this system encodes the multimodal language signal as multiple channels that are hierarchically structured and coordinated over time. While this design feature and others have been prompted by the unique requirements of generating a sign language, these technologies have applications for the machine translation of written languages, the representation of other multimodal language signals, and the production of meaningful gestures by other animated virtual human characters. 
To evaluate the functionality and scalability of the most novel portion of this English-to-ASL MT design, this project has implemented a prototype-version of the planning-based classifier predicate generator. The classifier predicate animations produced by the system have been shown to native ASL signers to evaluate the output.

Generating american sign language classifier predicates for english-to-asl machine translation

Users of sign languages are often forced to use a language in which they have reduced competence simply because documentation in their preferred format is not available. While some research exists on translating between natural and sign languages, we present here what we believe to be the first attempt to tackle this problem using an example-based (EBMT) approach.
Having obtained a set of English–Dutch Sign Language examples, we employ an approach to EBMT using the ‘Marker Hypothesis’ (Green, 1979), analogous to the successful system of (Way & Gough, 2003), (Gough & Way, 2004a) and (Gough & Way, 2004b). In a set of experiments, we show that
encouragingly good translation quality may be obtained using such an approach.

/pdf/an-example-based-approach-to-translating-sign-language-3wtw6nqx9f.pdf

An Example-Based Approach to Translating Sign Language

Systems that automatically process sign language rely on appropriate data. We therefore present the ATIS sign language corpus that is based on the domain of air travel information. It is available for five languages, English, German, Irish sign language, German sign language and South African sign language. The corpus can be used for different tasks like automatic statistical translation and automatic sign language recognition and it allows the specific modelling of spatial references in signing space.

/pdf/the-atis-sign-language-corpus-40g5ebdfvy.pdf

The ATIS sign language corpus

The South African Sign Language Machine Translation System (SASL-MT System) takes as its input English text, and outputs an avatar signing the equivalent SASL. This paper describes our experiences to date with the implementation of a signing avatar, in the light of the specific requirements of Sign Language.

South African Sign Language Machine Translation System

We consider the modelling of a particular layout optimisation problem with cellular automata, namely, the LEGO construction problem. We show that this problem can be modelled easily with cellular automata, provided that cells are considered as clusters which can merge or split during each time step of the evolution of the cellular automaton. The LEGO construction problem has previously been solved with optimisation techniques based on simulated annealing and with a beam search approach, but we show that the use of cellular automata gives comparable results in general, and improves the results in many respects.

Cellular automata with cell clustering.

We describe the South African Sign Language Machine Translation project, and point out the role that the project is playing in the larger context of South African Sign Language and accessibility for the South African Deaf community.

South African sign language machine translation project

We report on the South African Sign Language Machine Translation project, and more specifically on our experience in extending a synchronous tree-adjoining grammar parser approach in order to generate non-manual signs and construct a suitable signing space. We show that post-processing of the target language tree, after transfer rules have been applied, results in a simple and efficient mechanism to generate information on non-manual signs for use in a signing avatar.

Lynette van Zijl

Papers

The ATIS sign language corpus

South African Sign Language Machine Translation System

Cellular automata with cell clustering.

South African sign language machine translation project

The South African sign language machine translation project: issues on non-manual sign generation