Using infrared to improve face recognition of individuals with highly pigmented skin

Face recognition is widely used for security and access control. Its performance is limited when working with highly pigmented skin tones due to training bias caused by the under-representation of darker-skinned individuals in existing datasets and the fact that darker skin absorbs more light and therefore reflects less discernible detail in the visible spectrum. To improve performance, this work incorporated the infrared (IR) spectrum, which is perceived by electronic sensors. We augmented existing datasets with images of highly pigmented individuals captured using the visible, IR, and full spectra and fine-tuned existing face recognition systems to compare the performance of these three. We found a marked improvement in accuracy and AUC values of the receiver operating characteristic (ROC) curves when including the IR spectrum, increasing performance from 97.5% to 99.0% for highly pigmented faces. Different facial orientations and narrow cropping also improved performance, and the nose region was the most important feature for recognition.

Development of database structure and indexing for siddha medicine system - A platform for siddha literature analytics

Background

Siddha Medicine system is one among the oldest traditional systems of medicines in India and has its entire literature in the Tamil language in the form of poems (padal in tamil). Even if the siddha poems are available in public domain, they are not known to other parts of the world because, researchers of other languages are not able to understand the contents of these poems and there exists a language barrier. Hence there is a need to develop a system to extract structured information from these texts to facilitate searching, comparing, analysis and implementing.

Objective

This study aimed at creating a comprehensive digital database system that systematically stores information from classical Siddha poems and to develop a web portal to facilitate information retrieval for comparative and logical analysis of Siddha content.

Methods

We developed an expert system for siddha (eSS) that can collect, annotate classical siddha text, and visualizes the pattern in siddha medical prescriptions (Siddha Formulations) that can be useful for exploration in this system using modern techniques like machine learning and artificial learning. eSS has the following three aspects: (1) extracting data from Siddha classical text (2) defining the annotation method and (3) visualizing the patterns in the medical prescriptions based on multiple factors mentioned in the Siddha system of medicine. The data from three books were extracted, annotated and integrated into the developed eSS database. The annotations were used for analyzing the pattern in the drug prescriptions as a pilot work.

Results

Overall, 110 medicinal preparations from 2 Siddhars (Agathiyar and Theran) were extracted and annotated. The generated annotations were indexed into the data repository created in eSS. The system can compare and visualize individual and multiple prescriptions to generate a hypothesis for siddha practitioners and researchers.

Conclusions

We propose an eSS framework using standard siddha terminologies created by WHO to have a standard expert system for siddha. This proof-of-concept work demonstrated that the database can effectively process and visualize data from siddha formulations which can help students, researchers from siddha and other various fields to expand their research in herbal medicines.

Opioid2MME: Standardizing opioid prescriptions to morphine milligram equivalents from electronic health records

BACKGROUND

The national increase in opioid use and misuse has become a public health crisis in the U.S. To tackle this crisis, the systematic evaluation and monitoring of opioid prescribing patterns is necessary. Thus, opioid prescriptions from electronic health records (EHRs) must be standardized to morphine milligram equivalent (MME) to facilitate monitoring and surveillance. While most studies report MMEs to describe opioid prescribing patterns, there is a lack of transparency regarding their data pre-processing and conversion processes for replication or comparison purposes.

METHODS

In this work, we developed Opioid2MME, a SQL-based open-source framework, to convert opioid prescriptions to MMEs using EHR prescription data. The MME conversions were validated internally using F-measures through manual chart review; were compared with two existing tools, as MedEx and MedXN; and the framework was tested in an external academic EHR system.

RESULTS

We identified 232,913 prescriptions for 49,060 unique patients in the EHRs, 2008-2019. We manually annotated a sample of prescriptions to assess the performance of the framework. The internal evaluation for medication information extraction achieved F-measures from 0.98 to 1.00 for each piece of the extracted information, outperforming MedEx and MedXN (F-Scores 0.98 and 0.94, respectively). MME values in the internal EHR system obtained a F-measure of 0.97 and identified 3% of the data as outliers and 7% missing values. The MME conversion in the external EHR system obtained 78.3% agreement between the MME values obtained with the development site.

CONCLUSIONS

The results demonstrated that the framework is replicable and capable of converting opioid prescriptions to MMEs across different medical institutions. In summary, this work sets the groundwork for the systematic evaluation and monitoring of opioid prescribing patterns across healthcare systems.

MESOCOSM: A mesocosm database management system for environmental nanosafety

Engineered nanomaterials (ENMs) are intentionally designed and produced by humans to revolutionize the manufacturing sector, such as electronic goods, paints, tires, clothes, cosmetic products, and biomedicine. With the spread of these ENMs in our daily lives, scientific research have generated a huge amount of data related to their potential impacts on human and environment health. To date, these data are gathered in databases mainly focused on the (eco)toxicity and occupational exposure to ENMs. These databases are therefore not suitable to build well-informed environmental exposure scenarios covering the life cycle of ENMs. In this paper, we report the construction of one of the first centralized mesocosm database management system for environmental nanosafety (called MESOCOSM) containing experimental data collected from mesocosm experiments suited for understanding and quantifying both the environmental hazard and exposure. The database, which is publicly available through https://aliayadi.github.io/MESOCOSM-database/, contains 5200 entities covering tens of unique experiments investigating Ag, CeO₂, CuO, TiO₂-based ENMs as well as nano-enabled products. These entities are divided into different groups i.e. physicochemical properties of ENMS, environmental, exposure and hazard endpoints, and other general information about the mesocosm testing, resulting in more than forty parameters in the database. The MESOCOSM database is equipped with a powerful application, consisting of a graphical user interface (GUI), allowing users to manage and search data using complex queries without relying on programmers. MESOCOSM aims to predict and explain ENMs behavior and fate in different ecosystems as well as their potential impacts on the environment at different stages of the nanoproducts lifecycle. MESOCOSM is expected to benefit the nanosafety community by providing a continuous source of critical information and additional characterization factors for predicting ENMs interactions with the environment and their risks.