Melanoma classification using efficientnets and ensemble of models with different input resolution
- Authors: Karki, Sagar , Kulkarni, Pradnya , Stranieri, Andrew
- Date: 2021
- Type: Text , Conference paper
- Relation: 2021 Australasian Computer Science Week Multiconference, ACSW 2021, Virtual, Online, 1-5 February 2021, ACM International Conference Proceeding Series
- Full Text:
- Reviewed:
- Description: Early and accurate detection of melanoma with data analytics can make treatment more effective. This paper proposes a method to classify melanoma cases using deep learning on dermoscopic images. The method demonstrates that heavy augmentation during training and testing produces promising results and warrants further research. The proposed method has been evaluated on the SIIM-ISIC Melanoma Classification 2020 dataset and the best ensemble model achieved 0.9411 area under the ROC curve on hold out test data. © 2021 ACM.
Non-invasive smartphone use monitoring to assess cognitive impairment
- Authors: Thang, Nguyen , Oatley, Giles , Stranieri, Andrew , Walker, Darren
- Date: 2021
- Type: Text , Conference paper
- Relation: 13th International Conference on Computer and Automation Engineering, ICCAE 2021 p. 64-67
- Full Text: false
- Reviewed:
- Description: Background: There are many tests for the early detection of Mild Cognitive Impairment (MCI) to prevent or delay the development of dementia, particularly amongst the elderly. However, many tests are complex and are required to be performed repeatedly. Cognitive assessment apps for a smartphone have emerged, but like other tests, they require the user to perform complex tasks like drawing time on a clock. Few studies have explored non-invasive ways of tracking and assessing MCI without having the user perform specific tests. Objective: This research ultimately aims to develop an app that runs in the background and collects smartphone usage data that correlates well with MCI test results. The focus of this preliminary study was to develop an app that collects usage data and common MCI questionnaires to see if usage data between people varied, and to establish associations between phone usage and cognitive tests results. Method: An android application was developed to gather data over three weeks by three volunteers (authors). Usage data collected included: SMS and call log, accelerometer, location, app usage, self-report. Cognitive tests implemented were Stroop, Go/No Go tests and absent-mindedness questionnaires. Due to the small sample size and Covid-19 restrictions (October 2020), location data was not reliable. SMS text was not collected for privacy reasons. Results: App categories can differentiate people, but the app usage cannot be used to distinguish people. © 2021 IEEE.
On the value of social media in health care
- Authors: Ukoha, Chukwuma , Stranieri, Andrew
- Date: 2021
- Type: Text , Journal article
- Relation: Journal of technology in behavioral science Vol. 6, no. 2 (2021), p. 419-426
- Full Text: false
- Reviewed:
- Description: The advent and diffusion of modern technologies have triggered the widespread adoption of social media by hospitals and medical clinics. Despite the increasing use of social media, its use cases in health care settings and the value proposition of each use case are yet to be explicated. To address this issue, this qualitative study explores the value of social media in health care. Relevant data were collected through semi-structured interviews with participants at 11 Australian hospitals and medical clinics. Common themes expressed by participants were identified through a thematic analysis of the transcripts. The findings revealed nine use cases of social media in health care: engaging in professional networking, harnessing patient feedback, promoting public health, educating professionals, educating patients, engaging with the public, crowdsourcing, conducting research, and patient collaboration. Furthermore, this study found that hospitals and medical clinics are not passive users of social media; rather, they make conscious decisions regarding whether, when, and how to use social media. Although social media can likely support various activities in health care settings, its value proposition for hospitals and medical clinics varies depending on the use case. Understanding such use cases and the value proposition in each use case will help more hospitals and medical clinics to incorporate social media strategically.
Open banking and electronic health records
- Authors: Stranieri, Andrew , McInnes, Angelique , Hashmi, Mustafa , Sahama, Tony
- Date: 2021
- Type: Text , Conference paper
- Relation: 2021 Australasian Computer Science Week Multiconference, ACSW 2021, Virtual, Online, 1-5 February 2021, ACM International Conference Proceeding Series
- Full Text:
- Reviewed:
- Description: The Open Banking model is a data sharing model emerging in financial services sector that involves technological and regulatory innovations designed to facilitate access to banking records by third party providers such as payment service providers. The model is predicted to disrupt financial services and encourage a wave of new third-party providers offering innovative services that will change the relationship between customers and banks. This article juxtaposes the Open Banking model against models for electronic health records. Providers that could supply innovative third party services with health record data if an Open Banking model were adopted in the health care industry are advanced. © 2021 ACM.
Rethinking IS Graduates Work-readiness: Employers' perspectives
- Authors: Faisal, Nadia , Chadhar, Mehmood , Goriss-Hunter, Anitra , Stranieri, Andrew
- Date: 2021
- Type: Text , Conference paper
- Relation: 27th Annual Americas Conference on Information Systems (AMCIS)
- Full Text: false
- Reviewed:
- Description: Being a significant stakeholder in the graduates' employment outcomes, it is vital to understand employers' perceptions of graduates' work-readiness. However, existing information systems (IS) literature focuses mainly on the perceptions of students or universities. This paper aims to fill this gap by analysing scoping interviews conducted with graduate recruiters and industry experts in Australia regarding attributes that can improve graduates' employment prospects in the information and communication technology industry. A preliminary investigation based on grounded theory identified three emergent themes from the data: behaviors, skills, and knowledge levels. Based on the findings, this study proposes an IS graduate work-readiness framework that can help universities to develop academic programs aimed at enhancing desirable skills and attitudes among IS graduates' employment.
The design of a smartbrush oral health installation for aged care centres in Australia
- Authors: Grzegorz Broda, Lukasz , Oseni, Taiwo , Stranieri, Andrew , Marino, Rodrigo , Robinson, Jodie , Yates, Mark
- Date: 2021
- Type: Text , Conference paper
- Relation: 5th International Conference on Medical and Health Informatics, ICMHI 2021 p. 176-180, virtual online, 14-16 May 2021, ICMHI '21: Proceedings of the 5th International Conference on Medical and Health Informatics
- Full Text: false
- Reviewed:
- Description: The oral health of residents in aged care centres in Australia is poor, contributing to infections, hospital admissions, and increased suffering. Although the use of electric toothbrushes has been deployed in many centres, smartbrushes that record and transmit information about brushing patterns and duration are not routinely deployed. Yet, the use of smartbrushes for aged care residents promises better oral care. Thus, a study aimed at investigating the appropriateness and suitability of a smartbrush for aged care residents is currently underway. Due to the peculiarity of the aged care setting, the incorporation of smartbrushes into residents' care does require careful planning and design considerations. This paper describes an initial design process undertaken through the use of an actor to understand the important elements to be incorporated whilst installing a smartbrush for use in aged care settings. The design covers configuration settings of the brush and app, including ergonomic factors related to brush and smartphone placement. A design science approach led to an installation re-design and a revised protocol for the planned study, the ultimate aim being to design installations to enhance perceived usefulness, ease of use, and attitudes towards the incorporation of smartbrushes for improving oral health care for aged care residents. © 2021 ACM.
Understanding the gap between academics and game developers : an analysis of gamasutra blogs
- Authors: Greenwood, Jordan , Achterbosch, Leigh , Stranieri, Andrew , Meredith, Grang
- Date: 2021
- Type: Conference paper
- Relation: 15th International Conference on Interfaces and Human Computer Interaction, IHCI 2021 and 14th International Conference on Game and Entertainment Technologies, GET 2021 - Held at the 15th Multi-Conference on Computer Science and Information Systems, MCCSIS 2021 p. 141-148
- Full Text: false
- Reviewed:
- Description: Communication between industry and academia in the fields of game development to date has been limited to the detriment of both groups. This lack of communication is more commonly known as the academia-industry divide. Blogs have been advanced as a good medium to reduce the divide because they allow academics to practice presenting their knowledge to different audiences and allows them to get suggestions and feedback from game developers. This study analyzed all blogs posted by members of Gamasutra.com in a 13-month span between March 2020 and April 2021. Forty-four of the 767 blogs were found to be referencing academic sources. Using Walton and Krabbe’s dialogue types, we discovered how academia was trying to communicate their knowledge relevant to game development and report on the extent to which academia had tried to influence game design and development. Results showed that the divide is real and that access to research information for the public is still quite difficult. The results also illustrate that academia only have had a small influence on the gaming industry and that only a small amount of gaming researchers were communicating their theories of game development. © MCCSIS 2021.All right reserved.
Blockchain leveraged decentralized IoT eHealth framework
- Authors: Uddin, Ashraf , Stranieri, Andrew , Gondal, Iqbal , Balasubramanian, Venki
- Date: 2020
- Type: Text , Journal article
- Relation: Internet of Things Vol. 9, no. March 2020 p. 100159
- Full Text:
- Reviewed:
- Description: Blockchain technologies recently emerging for eHealth, can facilitate a secure, decentral- ized and patient-driven, record management system. However, Blockchain technologies cannot accommodate the storage of data generated from IoT devices in remote patient management (RPM) settings as this application requires a fast consensus mechanism, care- ful management of keys and enhanced protocols for privacy. In this paper, we propose a Blockchain leveraged decentralized eHealth architecture which comprises three layers: (1) The Sensing layer –Body Area Sensor Networks include medical sensors typically on or in a patient body transmitting data to a smartphone. (2) The NEAR processing layer –Edge Networks consist of devices at one hop from data sensing IoT devices. (3) The FAR pro- cessing layer –Core Networks comprise Cloud or other high computing servers). A Patient Agent (PA) software replicated on the three layers processes medical data to ensure reli- able, secure and private communication. The PA executes a lightweight Blockchain consen- sus mechanism and utilizes a Blockchain leveraged task-offloading algorithm to ensure pa- tient’s privacy while outsourcing tasks. Performance analysis of the decentralized eHealth architecture has been conducted to demonstrate the feasibility of the system in the pro- cessing and storage of RPM data.
Dynamically recommending repositories for health data : a machine learning model
- Authors: Uddin, Md Ashraf , Stranieri, Andrew , Gondal, Iqbal , Balasubramanian, Venki
- Date: 2020
- Type: Text , Conference proceedings
- Relation: 2020 Australasian Computer Science Week Multiconference, ACSW 2020
- Full Text: false
- Reviewed:
- Description: Recently, a wide range of digital health record repositories has emerged. These include Electronic Health record managed by the government, Electronic Medical Record (EMR) managed by healthcare providers, Personal Health Record (PHR) managed directly by the patient and new Blockchain-based systems mainly managed by technologies. Health record repositories differ from one another on the level of security, privacy, and quality of services (QoS) they provide. Health data stored in these repositories also varies from patient to patient in sensitivity, and significance depending on medical, personal preference, and other factors. Decisions regarding which digital record repository is most appropriate for the storage of each data item at every point in time are complex and nuanced. The challenges are exacerbated with health data continuously streamed from wearable sensors. In this paper, we propose a recommendation model for health data storage that can accommodate patient preferences and make storage decisions rapidly, in real-time, even with streamed data. The model maps health data to be stored in the repositories. The mapping between health data features and characteristics of each repository is learned using a machine learning-based classifier mediated through clinical rules. Evaluation results demonstrate the model's feasibility. © 2020 ACM.
- Description: E1
Gestalt based evaluation of health information diagrams
- Authors: Sharma, Vishakha , Stranieri, Andrew , Burstein, Frada , Warren, Jim , Firmin, Sally
- Date: 2020
- Type: Text , Conference paper
- Relation: 24th International Conference Information Visualisation, IV 2020 Vol. 2020-September, p. 195-201
- Full Text: false
- Reviewed:
- Description: Diagrams for four different health care settings have been proposed: Snapshot Diagram, Diagnosis Diagram, Strength of Evidence Diagram and Patient Pathway Diagram The availability of large amount of digital health care data and potential to utilize its benefits led to the development of these diagrams. This paper presents an analysis of the diagrams based on the selection of a subset of Gestalt principles deemed relevant for each diagram. Although Gestalt and human-computer interaction principles are advanced to apply to all diagrams or user interfaces, in practice a sub-set of principles must be selected to evaluate a diagram or interface The selection of a subset of principles to use on a diagram has not been widely studied. This paper presents an approach for identifying a subset of relevant Gestalt principles tailored for each of the four diagrams advanced for health care settings. © 2020 IEEE.
Online dispute resolution in mediating EHR disputes : a case study on the impact of emotional intelligence
- Authors: Bellucci, Emilia , Venkatraman, Sitalakshmi , Stranieri, Andrew
- Date: 2020
- Type: Text , Journal article
- Relation: Behaviour and Information Technology Vol. 39, no. 10 (2020), p. 1124-1139
- Full Text:
- Reviewed:
- Description: An Electronic Health Record (EHR) is an individual’s record of all health events that enables critical information to be documented and shared electronically amongst health care providers and patients. The introduction of an EHR, particularly a patient-accessible EHR, can be expected to lead to an escalation of enquiries, complaints and ultimately, disputes. Prevailing opinion is that Online Dispute Resolution (ODR) systems can help with the mediation of certain types of disputes electronically, particularly systems which deploy Artificial Intelligence (AI) to reduce the need for a human mediator. However, disputes regarding health tend to invoke emotional responses from patients that may conceivably impact ODR efficacy. This raises an interesting question on the influence of emotional intelligence (EI) in the process of mediation. Using a phenomenological research methodology simulating doctor–patient disputes mediated with an AI Smart ODR system in place of a human mediator, we found an association between EI and the propensity for a participant to change their previously asserted claims. Our results indicate participants with lower EI tend to prolong resolution compared to those with higher EI. Future research include trialling larger scale ODR systems for specific cohorts of patients in the area of health related dispute resolution are advanced. © 2019 Informa UK Limited, trading as Taylor & Francis Group.
Rapid health data repository allocation using predictive machine learning
- Authors: Uddin, Ashraf , Stranieri, Andrew , Gondal, Iqbal , Balasubramanian, Venki
- Date: 2020
- Type: Text , Journal article
- Relation: Health Informatics Journal Vol. 26, no. 4 (2020), p. 3009-3036
- Full Text:
- Reviewed:
- Description: Health-related data is stored in a number of repositories that are managed and controlled by different entities. For instance, Electronic Health Records are usually administered by governments. Electronic Medical Records are typically controlled by health care providers, whereas Personal Health Records are managed directly by patients. Recently, Blockchain-based health record systems largely regulated by technology have emerged as another type of repository. Repositories for storing health data differ from one another based on cost, level of security and quality of performance. Not only has the type of repositories increased in recent years, but the quantum of health data to be stored has increased. For instance, the advent of wearable sensors that capture physiological signs has resulted in an exponential growth in digital health data. The increase in the types of repository and amount of data has driven a need for intelligent processes to select appropriate repositories as data is collected. However, the storage allocation decision is complex and nuanced. The challenges are exacerbated when health data are continuously streamed, as is the case with wearable sensors. Although patients are not always solely responsible for determining which repository should be used, they typically have some input into this decision. Patients can be expected to have idiosyncratic preferences regarding storage decisions depending on their unique contexts. In this paper, we propose a predictive model for the storage of health data that can meet patient needs and make storage decisions rapidly, in real-time, even with data streaming from wearable sensors. The model is built with a machine learning classifier that learns the mapping between characteristics of health data and features of storage repositories from a training set generated synthetically from correlations evident from small samples of experts. Results from the evaluation demonstrate the viability of the machine learning technique used. © The Author(s) 2020.
The delicate balance of communicational interests : a Bakhtinian view of social media in health care
- Authors: Ukoha, Chukwuma , Stranieri, Andrew
- Date: 2020
- Type: Text , Journal article
- Relation: Journal of Information, Communication and Ethics in Society Vol. 19, no. 2 (2020), p. 236-248
- Full Text: false
- Reviewed:
- Description: Purpose: This paper aims to use the writings of Mikhail Bakhtin to reveal new insights into the role and impact of social media in health-care settings. Design/methodology/approach: With the help of Bakhtin’s constructs of dialogism, polyphony, heteroglossia and carnival, the power and influences of the social media phenomenon in health-care settings, are explored. Findings: It is apparent from the in-depth analysis conducted that there is a delicate balance between the need to increase dialogue and the need to safeguard public health, in the use of social media for health-related communication. Bakhtin‘s constructs elucidate this delicate balance and highlight the need for health-care providers that use social media to find the right balance between these competing communicational priorities. Originality/value: This paper advances a nascent theoretical approach to social media research. By applying Bakhtinian ideas to consumer health informatics, this paper has the potential to open a new approach to theorizing the role of social software in health-care settings. Stakeholders in digital health will find this paper useful, as it opens up dialogue to further discuss the role of social media in health care. © 2020, Emerald Publishing Limited.
Towards smart online dispute resolution for medical disputes
- Authors: Bellucci, Emilia , Stranieri, Andrew , Venkatraman, Sitalakshmi
- Date: 2020
- Type: Text , Conference proceedings , Conference paper
- Relation: Proceedings of the Australasian Computer Science Week Multiconference (ACSW 2020); Melbourne, Australia; 3rd-7th February 2020. p. 1-5
- Full Text: false
- Reviewed:
- Description: With the advancements in technologies, digitization of health records in the healthcare industry is undertaking a rapid revolution. This is further fueled with the entrance of Internet of Things (IoT), where mobile health devices have resulted in an explosion of health data and increased accessibility via wireless communications and sensor networks. With the introduction of an Electronic Health Record (EHR) system as an important venture for the general health and wellbeing of a country's citizens, privacy issues and medical disputes are expected to rise. In addition to critical health information being documented and shared electronically, integrating data from diverse smart medical IoT devices are leading towards increasingly more complex disputes that require immense time and effort to resolve. Online dispute resolution (ODR) programs have been successfully applied to cost-effectively help disputants resolve commercial, insurance and other legal disputes, but as yet have not been applied to healthcare. This paper takes a modest step in this direction, firstly to identify the drivers of medical disputes that include patient empowerment and technology advancements and trends. Secondly, we explore dispute resolution models and identify the status and limitations of current ODR systems.
- Description: This work was funded by the University of Ballarat Deakin University Collaborative Fund. 160134
A Decentralized Patient Agent Controlled Blockchain for Remote Patient Monitoring
- Authors: Uddin, Ashraf , Stranieri, Andrew , Gondal, Iqbal , Balasubramanian, Venki
- Date: 2019
- Type: Text , Conference proceedings
- Relation: 15th International Conference on Wireless and Mobile Computing, Networking and Communications, WiMob 2019 Vol. 2019-October, p. 207-214
- Full Text: false
- Reviewed:
- Description: Blockchain emerging for healthcare provides a secure, decentralized and patient driven record management system. However, the storage of data generated from IoT devices in remote patient management applications requires a fast consensus mechanism. In this paper, we propose a lightweight consensus mechanism and a decentralized patient software agent to control a remote patient monitoring (RPM) system. The decentralized RPM architecture includes devices at three levels; 1) Body Area Sensor Network-medical sensors typically on or in patient's body transmitting data to a Smartphone, 2) Fog/Edge, and 3) Cloud. We propose that a Patient Agent(PA) software replicated on the Smartphone, Fog and Cloud servers processes medical data to ensure reliable, secure and private communication. Performance analysis has been conducted to demonstrate the feasibility of the proposed Blockchain leveraged, distributed Patient Agent controlled remote patient monitoring system. © 2019 IEEE.
- Description: E1
A lightweight blockchain based framework for underwater ioT
- Authors: Uddin, Md , Stranieri, Andrew , Gondal, Iqbal , Balasubramanian, Venki
- Date: 2019
- Type: Text , Journal article
- Relation: Electronics (Switzerland) Vol. 8, no. 12 (2019), p.
- Full Text:
- Reviewed:
- Description: The Internet of Things (IoT) has facilitated services without human intervention for a wide range of applications, including underwater monitoring, where sensors are located at various depths, and data must be transmitted to surface base stations for storage and processing. Ensuring that data transmitted across hierarchical sensor networks are kept secure and private without high computational cost remains a challenge. In this paper, we propose a multilevel sensor monitoring architecture. Our proposal includes a layer-based architecture consisting of Fog and Cloud elements to process and store and process the Internet of Underwater Things (IoUT) data securely with customized Blockchain technology. The secure routing of IoUT data through the hierarchical topology ensures the legitimacy of data sources. A security and performance analysis was performed to show that the architecture can collect data from IoUT devices in the monitoring region efficiently and securely. © 2020 by the authors. Licensee MDPI, Basel, Switzerland.
An efficient selective miner consensus protocol in blockchain oriented iot smart monitoring
- Authors: Uddin, Ashraf , Stranieri, Andrew , Gondal, Iqbal , Balasubramanian, Venki
- Date: 2019
- Type: Text , Conference proceedings , Conference paper
- Relation: 2019 IEEE International Conference on Industrial Technology, ICIT 2019; Melbourne; Australia; 13th-15th February 2019 Vol. 2019-February, p. 1135-1142
- Full Text:
- Reviewed:
- Description: Blockchains have been widely used in Internet of Things(IoT) applications including smart cities, smart home and smart governance to provide high levels of security and privacy. In this article, we advance a Blockchain based decentralized architecture for the storage of IoT data produced from smart home/cities. The architecture includes a secure communication protocol using a sign-encryption technique between power constrained IoT devices and a Gateway. The sign encryption also preserves privacy. We propose that a Software Agent executing on the Gateway selects a Miner node using performance parameters of Miners. Simulations demonstrate that the recommended Miner selection outperforms Proof of Works selection used in Bitcoin and Random Miner Selection.
- Description: Proceedings of the IEEE International Conference on Industrial Technology
Are ERP simulation games assisting students to be job-ready? An Australian universities’ perspective
- Authors: Faisal, Nadia , Chadhar, Mehmood , Goriss-Hunter, Anitra , Stranieri, Andrew
- Date: 2019
- Type: Text , Conference paper
- Relation: 30th Australiasian Conference on Information Systems (ACIS), 9-11 December 2019, Perth, Australia
- Full Text:
- Reviewed:
- Description: Deep and rapid changes in digital enterprise technology exceed the ability of traditional teaching methods to prepare students for challenges encountered in modern enterprises. Researchers proposed different pedagogical approaches to teach ERP (Enterprise Resource Planning) concepts such as ERPsim games to enhance students’ learning and job-readiness. Although the ERPsim studies verified the role of these games in enhancing students’ learning, whether these games contribute to student’s job readiness still needs to be explored. Using the mixed-method approach, this research-in-progress is designed to fill this gap by investigating the role of ERPsim game in increasing skills, learning levels, and job-readiness among university students in Australia. The findings from this study can contribute to the improvement of ERP pedagogical techniques. In addition, this research-in-progress will provide a concrete mapping to align learning outcomes/skills with ICT industry competencies standards as defined in SFIA (Skills framework for Information Age) and AQF (Australian Qualifications Framework).
Blockchain leveraged task migration in body area sensor networks
- Authors: Uddin, Ashraf , Stranieri, Andrew , Gondal, Iqbal , Balasubramanian, Venki
- Date: 2019
- Type: Text , Conference proceedings , Conference paper
- Relation: 25th Asia-Pacific Conference on Communications, APCC 2019 p. 177-184
- Full Text:
- Reviewed:
- Description: Blockchain technologies emerging for healthcare support secure health data sharing with greater interoperability among different heterogeneous systems. However, the collection and storage of data generated from Body Area Sensor Net-works(BASN) for migration to high processing power computing services requires an efficient BASN architecture. We present a decentralized BASN architecture that involves devices at three levels; 1) Body Area Sensor Network-medical sensors typically on or in patient's body transmitting data to a Smartphone, 2) Fog/Edge, and 3) Cloud. We propose that a Patient Agent(PA) replicated on the Smartphone, Fog and Cloud servers processes medical data and execute a task offloading algorithm by leveraging a Blockchain. Performance analysis is conducted to demonstrate the feasibility of the proposed Blockchain leveraged, distributed Patient Agent controlled BASN. © 2019 IEEE.
- Description: E1
Comparing Pixel N-grams and bag of visual word features for the classification of diabetic retinopathy
- Authors: Kulkarni, Pradnya , Stranieri, Andrew , Jelinek, Herbert
- Date: 2019
- Type: Text , Conference proceedings
- Relation: ACSW 2019: Australasian Computer Science Week 2019;Sydney NSW Australia; January 29 - 31, 2019; published in Proceedings of the Australasian Computer Science Week Multiconference p. 1-7
- Full Text: false
- Reviewed:
- Description: The extraction of Bag of Visual Words (BoVW) features from retinal images for automated classification has been shown to be effective but computationally expensive. Histogram and co-variance matrix features do not generally result in models that have the same predictive accuracy as BoVW and are still computationally expensive. The discovery of features that result in accurate image classification on computationally constrained devices such as smartphones would enable new and promising applications for image classification. For example, smartphone retinal cameras can conceivably make diabetic retinopathy widely available and potentially reduce undiagnosed retinopathy if it could be achieved with computationally simple classification algorithms. A novel image feature extraction technique inspired by N-grams in text mining, called 'Pixel N-grams' is described that can serve this purpose. Results on mammogram and texture classification have shown high accuracy despite the reduced computational complexity. However retinal scan classification results using Pixel N-grams lag behind BoVW approaches. An explanation for the relative poor performance of Pixel N-grams with diabetic retinopathy that draws on concepts associated with the No Free Lunch theorem are presented.