Medical Alert System

Road traffic accidents are being recognised as a major problem in developing countries. A mechanism to prevent these road accidents is the need of the hour. It is hoped that the mechanism which we introduced can prevent accidents occurring due to overtaking. In order to accomplish the task, we use a camera which is fixed on the rear view mirror of our vehicle which collects the footage of the opposite lane up to 100m. This footage is used to detect any obstacle present using real time computer vision. If an obstacle is present, driver will be alarmed so. This will ensure overtaking only at safe conditions

The purpose of this paper is to solve the road safety issues present in our country. In today's scenario lots of accidents happen on roads which result in a huge number of fatalities. The idea is to introduce a Vehicle to Vehicle communication protocol that will enable the vehicles moving on the road to communicate with each other in a wireless medium. The wireless communication between any two vehicles is possible using Radio Frequency (RF) communication with a range of 100 meters to 300 meters depending on the module and the antenna used. This Vehicle to vehicle communication also known as V2V is essential in areas with no internet connectivity or GSM (Global System for Mobile) based communication. This established connection between vehicles is used to communicate data like accident occurrence, vehicles ahead, vehicles coming from opposite lane and other warnings. The NRF module is used to establish the wireless Radio Frequency (RF) communication between vehicles. The NRF modules present in both the vehicles communicate with each other and send the message over a wireless medium. This type of communication is much more reliable than internet or mobile based communication. All the functions are controlled by an Arduino Uno R3 microcontroller which integrates all the sensors present. This technology prevents accidents on roads which is one of the most major issues in our country causing loss of lives. The connected vehicles on the roads provide smooth traffic, tackling congestion and in case of any human error the vehicle can communicate with the other vehicle in a quick and wireless medium cautioning the other driver thereby avoiding fatalities.

The neonatal incubator is an apparatus that provides a closed and controlled environment for the sustenance of premature babies. But recently, many premature babies have lost their lives due to lack of proper monitoring of the incubator that leads to accidents (leakage of gas and overheating causing short circuits and eventually, the bursting of incubators). This project deals with the cost- effective design of an embedded device that monitors certain parameters such as pulse rate of the baby, temperature, humidity, essential amount of gas and light inside the incubator. The details are sent as a message to the doctor or nurse through IoT, so that proper actions can be taken in advance, to maintain the apt environment inside the incubator and ensure safety to the infant's life. So, the objective of this project is to overcome the above-mentioned drawbacks and provide a safe and affordable mechanism for monitoring the incubator.

Numerous cases of missing are reported in Kerala. Our local department has the mechanism to watch every point on city. But it's not effectively helpful to seek out the missing person. It is found difficult to seek out an individual from crowd. During this paper, we present a mechanism to seek out the missing person or object from crowd by using opencv and image processing. The image of the person is subjected to image processing, which is taken as input for tracking the person or object from crowd using opencv. When the person or object is detected from video of crowd, an alert system is employed as a reminder. When an alert signal is produced, the authority can take necessary action within a short time. The objects are detected using one shot learning in real time scenarios. Main objective of the one shot learning is to detect various objects in real time video sequences. Once the network is trained by employing a single image of an individual, then the trained model can recognize that person in any angle. This real time analysis of the environment enables security and surveillance. This model can be implemented in different surveillance devices to detect attacks on many places like schools, government offices and hospitals where arms are completely restricted. I. INTRODUCTION Different mechanisms are used by our department of local government to monitor each purpose on town. But it's not effectively useful to find the missing person from crowd. We have a tendency to gift a mechanism to find the missing person from crowd by victimisation one shot learning. Main objective of the one shot learning is to discover various objects in real time video sequences. Once the network is trained by employing a single image of someone, then the trained model will acknowledge that person in any angle. This real time analysis of the environment allows security and surveillance. This model can be implemented in different police investigation devices to discover attacks on several places like faculties, government offices and hospitals wherever arms are completely restricted.

52th Congress on general and family medicine 52. Kongress für Allgemeinmedizin und Familienmedizin: Deutsche Gesellschaft für Allgemeinmedizin und Familienmedizin (German College of General Practitioners and Family Physicians) 9/13 - 9/15/2018 Innsbruck/Austria

Message clé          : Le potentiel de l’ordinateur pour l’amélioration la prescription par le biais des alertes automatiques n’est que très insuffisamment exploité.
Soumission antérieure : Non
Contenu              : Contexte et objectifs :
La 11e convention incite à l’utilisation d’un logiciel métier créant des alertes lors de la prescription. Quelle est la pertinence de ces alertes ?

Méthode utilisée :
À ce jour, j’ai évalué pour 32 patients choisi au hasard les alertes automatiques lors de la prescription de leur traitement habituel. 194 médicaments prescrits généraient 677 alertes de tous niveaux. 80 alertes étaient des doublons et 194 dû à l’insuffisance de la base de données médicamenteuse. 59 alertes interprétaient mal une pathologie saisie en mode structuré, par exemple (e.g.)  « hémochromatose » a été interprété comme « cirrhose/insuffisance du foie ».
16 alertes étaient évitables en prenant en compte l’âge du patient (e.g. statines et triptanes affichent « contre-indication absolue chez la femme en âge de procréer »).
26 ne considéraient pas une co médication protectrice (e.g. anti ulcéreux ou antiépileptique) et 18 pas la non simultanéité de la prise.  La prise en compte de la tension artérielle aurait évité 23 alertes, celle du dosage 19 (e.g. Aspirine en « dose cardiaque »), dont certaines étaient grossièrement fausses (e.g. alerte pour l’association d’un médicament dépresseur du système nerveux central avec des sprays d’asthme ou de nitrate contenant des traces d’alcool). La prise en compte de la clairance, de l’HbA1c et du TSH aurait évitée respectivement 95, 9, 2 alertes. 108 étaient manifestement infondées, mais je n’ai pas compris la cause (e.g. un IEC affichait « contre-indiqué chez l’hypertendu» ou la première prescription d’insuline auprès d’un diabétique type I sans traitement antérieur affichait « contre-indiqué car association avec pioglitazone »). J’ai évalué 147 alertes comme inévitables (21.6.%), mais dans aucun cas l’alerte ne m’a fait changer le traitement.
Jusqu’au congrès, une centaine de traitements seront évalués.


Principaux résultats :
Les alertes automatiques lors de la prescription sont inutiles voire inutilisables.

Conclusion et/ou question à débattre :
Le potentiel de l’ordinateur pour l’amélioration de la prescription n’est pas exploité.