In this paper, we report on the discovery process that resulted from a series of semioparticipatory workshops with Deaf community codesigners. This process was part of the user research conducted to substantiate the design of solutions to support autonomy in communication and information access for deaf persons who are sign language users. In our user research, we carried out four semio-participatory workshops, as defined by the socially aware design approach. We invited members of a Sign Language community into a democratic design process, providing interested parties with opportunities to reflect on experiences and preferences in the context of automatic sign language processing (ASLP) systems. Our main contribution is the formulation of 63 socio-technical good practices for the design of ASLP systems organized at the social, pragmatic, semantic and syntactic levels for both their human and technological aspects. Two other contributions resulted from our literature review and workshop planning: Firstly, we formalized the steps necessary to engage with a deaf community in the codesign process. Secondly, we present an analysis of our research through the lenses of five calls to action: including Sign Language community members as codesigners, discussing real-world applications, broadening the concept of user interface guidelines to socio-technical good practices, identifying the challenges to find representative datasets, and discussing issues involved in standardized annotated videos in sign language. We thus present both empirical and methodological contributions to the field of Human-Computer Interaction.
In this paper we present procedures and outcomes from conducting four workshops, following the
Socially Aware Design approach [
Our recent work [
2023 Copyright for this paper by its authors. Use permitted under Creative Commons License Attribution 4.0 International (CC BY 4.0).
interested parties as part of the research team, in a collaborative perspective towards a universal design,
including Sign Language Community members and hearing persons with diverse backgrounds. In this
paper, we use the term Deaf Community or Sign Language Community to refer to a subpopulation
among the diverse and larger group of people who are D/deaf [
With this human-centered approach as our research core, we have been conducting work on the
design of technology for, with and by a Deaf Community that builds upon the notions of Socially Aware
Design. Baranauskas [
Our work with a Deaf Community has progressed through several stages, focusing mainly on the
design of ASLP systems. As a starting point for the codesign of such interactive systems, we conducted
a systematic literature review of user studies for the design of ASLP systems [
In this paper, we share our understanding of socio-technical aspects that are involved in the design
of ASLP systems with a Deaf community as codesigners. In order to uncover evidence on
sociotechnical aspects, we relied on the guidance of the socially aware design approach by conducting four
semio-participatory workshops with Deaf community codesigners. This discovery process required
paying attention to details of what the codesign sessions with deaf signers revealed to us. Our main
contribution is the formulation of socio-technical good practices for the design of ASLP systems.
Additionally, we present two other contributions that resulted from conducting our literature review
and planning each workshop, as we realized they could help other researchers who carry out similar
work (e.g., with deaf communities or in the design and development of ASLP systems). The first
contribution is an analysis of various aspects of our research through the lenses of the five calls to action
proposed by [
Socially Aware Design is a human-computer interaction approach, proposed by Baranauskas [
The semio-participatory framework, a representation that is helpful in explaining Socially Aware
Design, considers society, or a sample of it, as a wrapper for three levels of culture study, as layers of
a semiotic onion [
Inclusive participatory practices involve participatory sessions with interested parties, providing
communication support, a physically accessible environment, and easy-to-use artifacts. Three concrete
artifacts were used during the semio-participatory workshops in this research: the Stakeholders
Identification Diagram, the Evaluation Frame and the Semiotic Framework [
The Stakeholders Identification Diagram artifact, used in Semio-Participatory Workshop 1, is a graphical representation consisting of five concentric circles: Starting from the center, a circle represents Operation (intended solution), followed by Contribution (main actors and responsible parties), Source (clients and suppliers), Market (partners and competitors) and Community (bystanders and legislators). This means that stakeholders who are closer to the Operation are those who can collaborate the most with the project. This artifact helps participants in identifying interested parties, from the four above mentioned categories, who they believe would be key to participate in the codesign process.
The Evaluation Frame artifact, used in Semio-Participatory Workshops 3 and 4, supports brainstorming sessions, where codesigners socialize questions and problems, as well as ideas and solutions for the technology’s design, taking into account each actor in the Stakeholders Identification Diagram artifact. For this research, we proposed an adaptation of the Evaluation Frame artifact, which presents each identified stakeholder as well as a standard representation for questions and problems, and ideas and solutions. Each element includes short and simplified texts side by side with a representative image in a separate sheet shown one at a time for reducing memory overload.
The Semiotic Framework (or Semiotic Ladder) [[
A significant number of related works that have applied concepts and artifacts of Socially Aware Design and conducted semio-participatory workshops have been reported in the literature. Some have studied meta-communication in inclusive scenarios [23] or proposed to extend the approach to include cultural and value perspectives to the design of interactive systems [24], whereas other have presented frameworks for assistive technology design [[25], [26], [27], [28]].
Since we conducted the field research in Mexico, we felt motivated to learn about characteristics of local research on ASLP systems and types of user studies related to this topic. We were interested to know whether there were research groups working on this topic, to understand research characteristics and to confirm whether investigation has been conducted on user studies on this domain.
In this exploratory study, using Google Scholar we found 43 works (14 in Spanish and 29 in English) with publication years between 2002 and 2019. Nineteen works were published in 2016 and 2017, the period of most interest in this topic in the country.
The works we found were conducted in 25 institutions from eleven states and Mexico City (Ciudad de México, CDMX). Some were carried out in collaboration with institutions from different states or countries (such as the United States and Italy). Five states stand out with higher numbers of scientific production on ASLP systems: CDMX (15), State of Mexico (6), Oaxaca (6), Puebla (6), and Veracruz (6), whereas five institutions lead this production: Instituto Politécnico Nacional (IPN, 11), Universidad Autónoma del Estado de México (UAEM, 8), Universidad Tecnológica de la Mixteca (UTM, 6), Benemérita Universidad Autónoma de Puebla (BUAP, 4) and Universidad Veracruzana (UV, 4).
Twenty-four papers come from eight research groups in six institutions, reporting advances of their work. Six papers from UTM in Oaxaca, eight papers from three different groups of IPN in CDMX, five papers from two groups of UAEM in México state (Teotihuacan Valley and Texcoco, respectively), two papers from Universidad Tecnológica de Puebla (UTP) and BUAP, and three works from UV. Based on the number of dedication years and number of publications, we can infer that studies on ASLP systems in Mexico are stronger in these groups. Nineteen are isolated works from different research groups with only one publication each.
Among the 43 works we found, thirty-seven reported research conducted on Mexican Sign Language (LSM, Lengua de Señas Mexicana), three on American Sign Language (ASL) and three did not specify a sign language. Regarding the research focus within sign language processing studies, fifteen worked merely on letters of the alphabet [[29], [30], [31], [32], [33], [34], [35], [36], [37], [38], [39], [40], [41], [42], six on letters and words [[18], [43], [44], [45], [46], [47]], four only on words [[48], [49], [50], [51]], four on sentences [[52], [53], [54], [55]], three on numbers and letters of the alphabet [[56], [57], [58]], two on numbers, letters and words [[59], [60]], two on words and sentences [[61], [62]], one on hand configuration and numbers [45], one on phonetic units [63], one on vowels [64] and four did not mention their focus [[65], [66], [67], [68]].
Considering that research on automatic sign language processing systems encompasses various
topics studied by many researchers around the world, we identified twenty-six works on ASLR systems,
thirteen on ASLT systems, two on ASLG associated with ASLR systems, one on a dictionary and one
on a database challenge contest. Amongst the ASLT systems group of papers, there are many with
ASLR systems characteristics. Therefore, there is a tendency in research on recognition of letters of the
alphabet, numbers, and isolated signs. This may “risk misrepresenting sign language recognition as a
gesture recognition problem, ignoring the complexity of sign languages as well as the broader social
context within which such systems must function” [
With respect to potential users’ involvement, in the related work (Section 3), thirty works reported some type of participation and thirteen did not mention anything about it. From works that mentioned user involvement, twenty-one noted that individuals participated in the process to compose the authors’ database, thirteen in tests of the proposed system, two in planning of the database, one in the needs, activities and context identification phase, and one in the opinion collection about the prototype design. Six works (20%) explicitly reported involvement of deaf persons as participants [[29], [33], [39], [52], [54], [60]]. Only one work (3,34%) has emphasis on the HCI field of research [33], where the author conducted the four phases of the HCI life cycle [69] including potential users as participants. Moreover, among these 20% works that are related to user studies, none neither investigated about context or society-centered design nor proposed socio-technical good practices for the design of ASLP systems.
In a previous study [
We conducted four Semio-participatory Workshops, which took place in a classroom at the local Association of the Deaf, as one-hour sessions every two weeks. The researchers (R1 and R2) were familiar with the Sign Language community since they were taking LSM lessons at the Association; however, they were not yet proficient enough to fully communicate with participants in LSM. Therefore, in all sessions we had support of a LSM interpreter, who also participated as codesigner. This interpreter helped recruit other members of the Sign Language community. The informed consent was one of the documents analyzed and approved, along the research project, by the university ethics committee for investigations conducted with human beings.
In the inclusive participatory practices, we used four types of artifacts. For the workshops, artifacts were printed, and participants could use colored markers, post-it notes, and stickers. Following the Socially Aware Design approach, the three artifacts we used are: the Stakeholders Identification Diagram, the Rating Scenarios, the Evaluation Frame, and the Semiotic Ladder. We filmed and transcribed the inclusive participatory practices.
In this section, we present participants and procedures, as well as outcomes from the four semioparticipatory workshops carried out biweekly for two months. Subsection 5.3 covers descriptions of two workshops since we ran them back-to-back using the same artifact. 5.1.
Semio-participatory Workshop 1 was conducted with five codesigners, in which participants from the Sign Language Community included two D/deaf persons, a mother of a D/deaf person, a LSM Interpreter and a Researcher. Three codesigners were women and two men, with age average of 38.6, ranging from 19 to 52 years old.
In our case of interest, the alternative solution is related to automatic sign language processing (ASLP) systems for sign language recognition, generation and translation. As the main activity of the Semio-participatory Workshop, we conducted the inclusive participatory practice using the original Stakeholders Identification Diagram artifact translated into Spanish and including ASLP systems as the Operation. In Operation (intended solution) is placed at the innermost and core circle of the stakeholders’ layers in the artifact, which range from the most (closest) to the least (farthest from the center) interested parties.
The task of the codesigners –members of the Sign Language community and researchers as a team— in this session was to identify the potential stakeholders. From the closest interested parties, in the Contribution category referring to main actors and responsible parties, through the Source category referring to clients and suppliers and the Market category referring to partners and competitors, to the Community category referring to bystanders and legislators.
As a deliverable of Semio-participatory Workshop 1, we produced a list of identified stakeholders for each category who codesigners inferred could collaborate in the Socially Aware Design process of ASLP systems. The four categories of the Stakeholders Identification Diagram artifact were elicited as follows: (a) Contribution, with nine stakeholders; (b) Source, with five stakeholders; (c) Market, with four stakeholders; and (d) Community, with two stakeholders.
In total, participants identified twenty categories of representatives as relevant stakeholders to take part as codesigners in the proposal of ASLP systems. Our participants belonged to four of those participant categories and make up a sufficiently diverse mix for a first iteration of codesign. 5.2.
Semio-participatory Workshop 2 was conducted two weeks after the first one. Out of seven codesigners, four participants were from the Sign Language Community–two deaf teachers, a mother, and a teacher of D/deaf persons–, a LSM Interpreter and two Researchers. Four codesigners were women and three men, with age average of 40.85, ranging from 20 to 57 years old.
In the first Semio-participatory workshop, we noticed a need to explore other possibilities of scenarios in which ASLP systems could be embedded, since at that moment we did not intend to codesign user interfaces. We designed fifteen scenarios (Table 1) in which ASLP systems could support communication and information access, reminding codesigners of the Stakeholders Identification artifact in each scenario.
In Semio-participatory Workshop 2, the Rating Scenarios artifact consisted in a set of previously designed and printed materials, presented in a random order, and numbered at the top left corner. Also, each scenario included a short text explanation next to an illustrative image at the center and, at the bottom, a 5-point Likert scale with smiley faces (ranging from Dislike very much to Like very much), which we called a like-scale. The Rating Scenarios artifact presents what has been proposed in literature for different types of ASLP systems, so as codesigners who were not aware of such systems could get a visual idea of possible related systems.
Displaying on the wall and explaining each scenario at a time, we invited participants to indicate, with a colored dot sticker, how much they liked the design alternative presented as the first task everyone should accomplish. As a second task, after discussing and using the like-scale for each individual scenario, we asked participants to number them from 1 to 15 according to their preferences - 1 being the one they preferred the most, and 15 the least - and discussing their motivations. For this second ranking, participants decided to form two groups, one with two D/deaf members and another with two hearing members. The interpreter decided not to participate in this second ranking task and researchers did not participate in any of the rating tasks.
As a deliverable of Semio-participatory Workshop 2, we generated a spreadsheet, which compiles scenario ratings from Task 1 (like-scale) and Task 2 (ordering). Among the 15 scenarios, six scenarios selected (#1, #2, #3, #6, #11 and #15) were taken for voting in the Workshop 3 to identify the two preferred solutions to design. 5.3.
Workshops 3 and 4 focused on the Evaluation Frame, in which, as prescribed by Socially Aware Design, we brainstormed and ranked questions and problems, and ideas and solutions for each stakeholder type.
Semio-participatory Workshops 3 and 4 were conducted two and four weeks after Workshop 2, respectively. Four codesigners participated in each workshop: two from the Sign Language Community–a deaf teacher and a teacher of D/deaf persons; as well as two D/deaf teachers, respectively–, a LSM Interpreter, and a Researcher. Three and two codesigners, respectively were women, with age average of 32 and 41 ranging from 20 to 44, and 32 and 52 years old, respectively.
The inclusive participatory practice began with the use of the Evaluation Frame artifact to carry out the brainstorming on questions and problems and ideas and solutions for the design of ASLP systems. Once the two scenarios were selected (#6 and #15), we conducted the session using an adaptation of the Evaluation Frame artifact. The artifact’s adaptation consisted of framing an illustrative image for each stakeholder identified in the Semio-participatory Workshop 1 next to its category and identification, and by its side two blank spaces with images and labels representing questions and problems, as well as ideas and solutions.
In Semio-participatory Workshop 3, scenario #15 was ranked as the first preferred scenario, with the argument that it is useful in a broad set of situations, from information access to large contents in written language to data exchange through diverse communication means (e.g., email, instant messaging). The scenario #6 was ranked as the second preferred scenario, for which researchers argued the motives for keeping it, since the main idea is to provide face-to-face bidirectional communication between signers and non-signers.
In order to analyze the Semio-participatory Workshops 3 and 4 data, we organized data collected during Workshop 1. This organization consisted in grouping together some of the potential interested parties into a more general type of stakeholder. Thus, from the original twenty stakeholder’s types, we converged into eighteen types. The deliverable of these two Workshops was a table with questions and problems, and ideas and solutions raised by codesigners, whose results we used to plot into the Semiotic Ladder artifact to organize the set of socio-technical good practices we have derived.
From Section 5 we can observe that the results from applying the artifacts (Stakeholders Identification Diagram, Rating Scenarios and Evaluation Frame) are linked, in the sense that a workshop depends on the results from a previous one. For instance, in Semio-participatory Workshop 1, the stakeholder “School teachers” was elicited as an actor who can contribute in the codesign of ASLP systems. In Semio-participatory Workshop 2, codesigners discussed the possible scenarios in light of the stakeholders elicited, so they were asked questions such as “Can you imagine if this scenario was available for your schoolteacher to work with you as a student? Would this be positive or negative?” to serve as a concrete example to reflect about. In Semio-participatory Workshop 3 and 4, codesigners brainstormed questions and problems, and ideas and solutions related to stakeholders identified in Semio-participatory Workshop 1 and to the two most preferred scenarios. Therefore, codesigners were invited to imagine what kind of questions and problems, and ideas and solutions could arise in a concrete scenario. For instance, a situation where a schoolteacher in a mainstream classroom with many mixed students (signers and non-signers) wants to conduct a pedagogical activity using the scenario #15 as a mediator for communication or collaboration between pairs of students.
As a result, we found a higher number of items on good practices (46) for the “Human information functions” (social, user experience and HCI aspects) than the number of items (17) for the “The IT platform” (technical aspects). Since we were more interested in the human and context aspects for the design of ASLP systems, this list of good practices is a contribution to this line of study that requires an interdisciplinary approach.
From top to bottom of the Semiotic ladder, firstly, we present the three “Human information functions:” Social, Pragmatic and Semantic levels.
1. Social level. Since nowadays still there are many misconceptions about universal sign language (SL), full literacy in written language, communication homogeneity, oralization, among others, the guideline “Educate people about Deaf culture and deaf people’s rights” is an important issue to address in many levels of the technologies design. From literature review in this paper and from a previous systematic review, we found that most of the research groups are creating their own sign language database, since there is not a public repository with standardized data for SL of each country. This leads us to the guideline “Have political support to make a distributed database an official location to share and to receive standardized data (annotated SL videos) as an open science repository to support inter-disciplinary research”. One concern of many in the Deaf community is they are contacted by many researchers when they need to substantiate their projects or to fulfil some work agenda. Unfortunately, when projects are “completed” or the workload is finished, they disappear. That is why we included the guideline “Understand who can be potential supporters in the general community to guarantee sustainability of the technologies’ adoption and maintenance”. Table 2 presents the recommended good practices for the Social level of the Semiotic Ladder Artifact. 2. Pragmatic level. The three good practices here meet perspectives of the above three social levels in a more practical way, respectively, “Invite users to switch roles concerning mode of communication”, “Populate the database with users’ collaboration”, and “Record/ register the entire conversation, translations”, since they can motivate people to know about Deaf culture, can support the data gathering for a national initiative, and can provide motivation for government or companies to sustain the infrastructure and services for the technology continuance of use. Table 3 presents the recommended good practices for the Pragmatic level of the Semiotic Ladder Artifact. 3. Semantic level. The guideline "Provide textual, audio and video tutorials" is to ensure a broad range of types of users can access and know how to use technologies. In many contexts of use, it is important to keep track of a conversation history, such as medical consultation, academic or lawyer advising; this leads to the guideline "Automatically create a conversation timeline to save translations". The guideline "Recommend usefulness of the result presented by the search or the translation" can support assessment of the technology use for its improvement. Table 4 presents the recommended good practices for the Semantic level of the Semiotic Ladder Artifact.
From top to bottom of the Semiotic ladder, secondly, we present the three “The IT platform”: Syntactic, Empirical and Physical levels.
4. Syntactic level. The guideline “Allow users to adjust accessibility settings” can be related to adjustments of color contrast, font size, video/ animation display size, SL to speech for interaction between deaf and blind persons, speed of information presentation, following the adequate standards. Many concerns towards data privacy in specific contexts of use were reported, in spite wanting to keep the record of translations to themselves; the ideal situation is having transparency and well-defined norms to “Provide different privacy protocols for users’ data depending on the facilities’ nature”. Table 5 presents the recommended good practices for the Syntactic level of the Semiotic Ladder Artifact. 1. Allow users to adjust accessibility settings 2. Ensure privacy of patients’ personal medical records 3. Provide different privacy protocols for users' data depending on the facilities’ nature 5. Empirical level. The guideline "Self-adapt SL recognition to users with multiple disabilities" regards to specificities of deaf signers who have another disability associated, in order to support positioning in the right place, avoid clicking the wrong icon, recognizing the adequate hand configuration, avoiding feedback that presses the user for quick action, among others. Since the ASLP systems could be embedded in many types of contexts of use, one concern was about the specialized vocabulary used, either to deaf signers who still have to collaboratively create with community new signs as they further achieve higher academic education in diverse areas of knowledge or to guarantee a robust and diverse SL database. For this, the guideline "Gather data from diverse specific' domains" was pro-posed. Table 6 presents the recommended good practices for the Empirical level of the Semiotic Ladder Artifact. 6. Physical level. The guideline "Design one large screen for all, or individual screens for each" attends to expectations for private and public technology use, as well as for individual and collective technology use of a text to SL translator system. As part of the technical planning for supporting both scenarios of technology chosen by participants, the guideline "Determine protocol and infrastructure for information storage" was included. Table 6 presents the recommended good practices for the Physical level of the Semiotic Ladder Artifact.
The primary goal of this research has been to share our understanding of socio-technical aspects that are involved in the design of ASLP systems with Deaf community members as codesigners. In order to uncover evidence on socio-technical aspects, we relied on the guidance of the socially aware design approach by conducting four semio-participatory workshops with Deaf community codesigners. By presenting how research was conducted, its results and socio-technical good practices we derived, we highlight the importance of considering not only potential users at the center of the design, but also their ecosystem (main actors, responsible parties, clients, suppliers, partners and competitors, bystanders and legislators) and the research impact of this broader view.
Our invitation for a Sign Language community to participate in this democratic design process resulted in an opportunity for all to reflect on and to share social and technical concerns regarding past experiences and personal preferences in the context of ASLP systems. Participants were more timid in the first workshop, and more participative in the next ones. They understood the potential benefits this type of technology can bring to their lives or to their children or relatives. One deaf participant reported twice she was having fun being a codesigner, she was learning from discussions, and perceiving her ideas were being valued by others and for the project.
We provided a set of sixty-three experience-grounded socio-technical good practices for the design
of ASLP systems from which we can follow to the next steps of the research. We recall that the two
scenarios chosen by codesigners were #6 (Real-time and in-person communication mediated by a glass
interface or a screen (signer and non-signer)) and #15 (Translation from text to SL). These good
practices mostly can be applied to both scenarios. However, the social level brings to the table more
concerns for Scenario 1, since for many stakeholders’ types it involves the conversation between two
persons in public environments. Some recurrent codesigner concerns had to do with the use of data
from translations versus privacy, educating people about Deaf culture and sign language, and learning
how researchers will deal with sign language specificities, such as domain specific language,
regionalisms, and providing other visual cues to facilitate understanding and to ensure a positive user
experience. These findings and good practices are not set in stone, as we understand the need to
complement them with other categories of stakeholders. However, we disclose this contribution to
invite whoever seeks to further investigate this subject in an interdisciplinary research team to broaden
their views to include the human and context aspects. Also, Artificial Intelligence researchers could
benefit from our set of good practices in order to discuss topics related to Fairness, Accountability,
Transparency, and Ethics (FATE) since relevant issues still need to be further addressed [
Looking back at the entire process, we share a discovery process in light of the Socially Aware Design approach. This discovery process is related to the way inclusive participatory practices can be conducted and artifacts can be adapted to promote participants’ engagement in the design of solutions with Deaf community codesigners. We found that an initial stimulus activity related to the topic to be discussed (e.g., presenting a video or an app, making a conversation with questions, inviting for a vote) can be ice-breaking while eliciting data. The core session with an artifact must not take too long. The choice of day and time as well as having a snack help keep energy at a high level. Artifacts that present the same information in different formats (e.g., short, simplified text and image) and are explained in the preferred language of the participants ensure inclusiveness. Sharing the workshop content beforehand with the interpreter can help him or her feel more relaxed and enjoy the activity. Colorful supplies, such as post-its and sharpies, different kinds of stickers (dots, smiley faces, numbers, thumbs up and down) can be seen superfluous, but participants get excited to choose colors and feel motivated to collaborate. Finally, comments and testimonies of participants are valuable, and they should be informed about this by asking them to write down their thoughts on post-its, and to further discuss in their preferred mode of communication.
Additionally, taking the five calls to action proposed by [
We had planned at least eight semio-participatory workshops in addition to the four described in this paper. Unfortunately, the COVID-19 pandemic made us rearrange our plans. We were not able to conduct remote workshops, since most participants come from low-income families, and they did not have infrastructure to continue. A remote continuation of the research was possible with Brazilian participants, which currently is an ongoing effort.
We thank the Doctoral Program in Educational Systems and Environments of BUAP, the Directorate for Special Education of the State’s Ministry of Education (SEP), “Casa del Sordo” and “CAM Jean Piaget,” all in Puebla, for their willingness to collaborate with our research. We thank all the volunteers for their time and for sharing their knowledge with the research team. The project is registered in Plataforma Brasil (CAAE) under No. 18708619.2.0000.8088.
[33] M. E. Gutiérrez-Martínez. 2018. Modelo de redes neuronales para el reconocimiento de señas en un contexto de levantamiento de una denuncia por robo. Tesis de Maestría en Sistemas Interactivos Centrados en el Usuario, Universidad Veracruzana, Xalapa, Veracruz. [34] B. Martínez-Seis, O. Pichardo-Lagunas, E. Rodriguez-Aguilar, E. Saucedo-Diaz. Identification of Static and Dynamic Signs of the Mexican Sign Language Alphabet for Smartphones using Deep Learning and Image Processing. Research in Computing Science, 148, 2019, 199-211. [35] E.M. Morales, O.V. Aparicio, P. Arguijo, R.Á. Armenta, A.H. López.. Traducción del lenguaje de señas usando visión por computadora. Research in Computing Science, 148, 2019, 79-89. [36] F. P. Pérez-Priego, J. M. Ceja-Olivares, J. F. Talamantes-Serrano, D. N. Rivera-Aguilar. Image recognition of Mexican Sign Language. In: Castillo Montiel, E.; Chimal Eguía, J.C.; Uriarte Arcia, A.; Cabrera Rivera, L. (Ed.). Research in Computing Science, vol. 58, pp 57-68, 2012, Retrieved August 13, 2020, from https://www.rcs.cic.ipn.mx/2012_58/RCS_58_2012.pdf [37] P. P. E. Rivas, O. Velarde-Anaya, S. González-López, P. P. Rivas, N. A. Álvarez-Torres.
Entrenamiento de una Red Neuronal para el Reconocimiento de Imágenes de Lengua de Señas Capturadas con Sensores Profundidad. Congr. Int. en Ing. Electrónica. Mem. ELECTRO, Vol. 39, pp. 55-59, 2017. Retrieved August 06, 2020, from https://arxiv.org/abs/1804.00508 [38] M. Rivera-Acosta, S. Ortega-Cisneros, J. Rivera, F. Sandoval-Ibarra. American Sign Language Alphabet Recognition Using a Neuromorphic Sensor and an Artificial Neural Network. Sensors 2017, 17, 2176; doi:10.3390/s17102176. [39] G. González-Saldaña, J. Sánchez-Cerezo, M. M. Díaz-Bustillo, A. Pérez-Ata. Recognition and
Classification of Sign Language for Spanish. Computación y Sistemas, Vol. 22, 1, 2018, 271–277. [40] F. Solís, C. Toxqui, D. Martínez. Mexican sign language recognition using Jacobi-Fourier moments. Engineering,07,700-705, 2015. DOI: 10.4236/eng.2015.710061. [41] F. Trujillo-Romero, F. E. Luis-Pérez, S. O. Caballero-Morales. Multimodal Interaction for Service Robot Control. 22nd International Conference on Electrical Communications and Computers (CONIELECOMP 2012), DOI: 10.1109/CONIELECOMP.2012.6189929. [42] F. Trujillo-Romero, S. O. Caballero-Morales. 3D Data Sensing for Hand Pose Recognition.
International Conference on Electronics, Communications and Computers (CONIELECOMP 2013), IEEE. DOI: 10.1109/CONIELECOMP.2013.6525769. [43] S. O. Caballero-Morales, F. Trujillo-Romero. 3D Modeling of the Mexican Sign Language for a
Speech-to-Sign Language System. Computación y Sistemas Vol. 17 No. 4, 2013 pp. 593-608. [44] C. R. Estrivero-Chavez, M. A. Contreras-Teran, J. A. Miranda-Hernandez, J. J. Cardenas-Cornejo, M. A. Ibarra-Manzano, D. L. Almanza-Ojeda. Toward a Mexican Sign Language System using Human Computer Interface. International Conference on Mechatronics, Electronics and Automotive Engineering (ICMEAE 2019), DOI: 10.1109/ICMEAE.2019.00010. [45] J. Ibarra-Leybón, M. del R. Barba-Ramírez, V. Picazo-Taboada. "SENSor Foto-Eléctrico Aplicado al Movimiento de los Dedos de las Manos". Computación y Sistemas Vol. 10, 1, 2006, 57-68. [46] L. M. Pérez, A. J. Rosales, F. J. Gallegos, A. V. Barba. LSM static signs recognition using image processing. 14th International Conference on Electrical Engineering, Computing Science and Automatic Control (CCE) Mexico City, Mexico, September 20-22, 2017. [47] F. Trujillo-Romero, S. O. Caballero-Morales. Towards the Development of a Mexican Speech-toSign-Language Translator for the Deaf Community. Acta Universitaria, vol. 22, marzo, 2012, pp. 83-89, Universidad de Guanajuato, Guanajuato, México, ISSN: 0188-6266. [48] J. Cervantes, F. García-Lamont, L. Rodríguez-Mazahua, A.Y. Rendon, A.L. Chau. Recognition of Mexican Sign Language from Frames in Video Sequences. In: Huang DS., Jo KH. (eds) Intelligent Computing Theories and Application. ICIC 2016. Lecture Notes in Computer Science, vol 9772.
Springer, Cham. https://doi.org/10.1007/978-3-319-42294-7_31. [49] G. García-Bautista, F. Trujillo-Romero, S. O. Caballero-Morales. Mexican Sign Language Recognition Using Kinect and Data Time Warping Algorithm. International Conference on Electronics, Communications and Computers (CONIELECOMP 2017), DOI: 10.1109/CONIELECOMP.2017.7891832. [50] I. Guyon, V. Athitsos, P. Jangyodsuk, H. J. Escalante. The ChaLearn gesture dataset (CGD 2011).
Machine Vision and Applications 25, 1929–1951 (2014).