Her safety: empowering women’s safety with hand sign based communication

International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395-0056

Volume: 13 Issue: 03 | mar 2026 www.irjet.net p-ISSN: 2395-0072

Her safety: empowering women’s safety with hand sign based communication

Mrs.

P. Yogashree1 , S. Mehareen2 , M. Rajalakshmi3 , S. Shamreen4

1Assistant Professor, Dept of Computer Science and Engineering Vivekanandha College of Engineering for Women, Tamilnadu, India

2UG Scholar, Dept of Computer Science and Engineering, Vivekanandha College of Engineering for Women, Tamilnadu, India

3UG Scholar, Dept of Computer Science and Engineering, Vivekanandha College of Engineering for Women, Tamilnadu, India

4UG Scholar, Dept of Computer Science and Engineering, Vivekanandha College of Engineering for Women, Tamilnadu, India

Abstract - Women's safety remains a critical concern, especially in situations where individuals are unable to communicate distress verbally. This paper presents HERSAFETY, an intelligent, contactless emergency alert system that uses hand gesture recognition for silent communication during dangerous situations. The system leverages computer vision and artificial intelligence to detect predefined hand gestures using a webcam in real time. A dual-stage gesture sequence consisting of a closed fist followed by a thumb fold is used to minimize false alarms. Upon detection, the system automatically sends an emergency email alert to a predefined contact using SMTP protocol. The system is implemented using OpenCV, MediaPipe, and Python, ensuring real-time performance without requiring specialized hardware. Experimental results demonstrate high detection accuracy and low response time, making the system a reliable and accessible solution for enhancing personal safety.

Key Words: Women's Safety, Gesture Recognition, Computer Vision, MediaPipe, Emergency Alert System, Artificial Intelligence

1. INTRODUCTION

Women's safety has become a major concern in modern society, especially in situations where individuals are unable to communicatedistresseffectively.Inmanyemergencyscenarios,victimsmaynothavetheopportunitytousemobiledevices or verballyaskforhelp.Traditionalsafetysystemssuchaspanicbuttons,mobileapplications,andvoice-basedassistantsrequire physicalinteractionoraudiblecommunication,whichmaynotbefeasibleincriticalsituations.Toaddresstheselimitations, intelligent and contactless safety solutions are required. With the rapid advancement of Artificial Intelligence (AI) and Computer Vision,gesture-basedcommunicationhasemergedasaneffectiveapproachforhuman-computerinteraction.These technologiesenablesystemstointerprethumanactionsinrealtimewithoutrequiringphysicalcontactorspeech.Computer visionframeworksallowcontinuousmonitoringofhandmovementsthroughcameras,whilemachinelearningalgorithmshelp inaccuratelyrecognizingpredefinedgestures.Suchsystemscanoperatesilentlyanddiscreetly,makingthemhighlysuitable foremergencysituationswheredrawingattentionmayincreaserisk.ThisprojectproposesHERSAFETY, anAI-basedhand gesturerecognitionsystemdesignedtoenhancewomen'ssafetythroughsilentcommunication.Thesystemcontinuously monitorshandgesturesusingawebcamandidentifiesapredefineddistresssignalbasedonatwo-stagegesturesequence. Upon detecting an emergency gesture, the system automatically sends an alert notification to a predefined contact using internet-based communication. By integrating real-time gesture detection, intelligent processing, and automated alert mechanisms,theproposedsystemimprovesresponsetimeandprovidesareliable,contactlesssolutionforpersonalsafety. Thisapproachenhancesaccessibility,ensuresprivacy,andoffersacost-effectivemethodforemergencycommunicationin criticalsituations

2. METHODOLOGY

Theproposedsystemisdesignedtoenhancewomen'ssafetybydetectingdistressgesturesandsendingreal-timealerts.The methodintegratescomputervision,artificialintelligence,andautomatedcommunicationtocreateanefficientandreliablesafety system. The system operates through several stages including video capture, gesture detection, validation, and alert transmission.

International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395-0056

Volume: 13 Issue: 03 | mar 2026 www.irjet.net p-ISSN: 2395-0072

2.1 System Design

Thesystemdesignfocusesondevelopinganintegratedarchitecturethatcombinesvideoprocessing,gesturerecognition,and communicationtechnologiestoenhancewomen'ssafety.Thesystemiscenteredaroundasoftware-basedprocessingunit, wherecomputervisionandartificialintelligencealgorithmsactasthemaincontrolcomponents. Thewebcam,processing modules,andcommunicationsystemworktogethertoenablecoordinatedsystemoperation.

Intheproposeddesign,thewebcamisusedtocapturereal-timevideoinputfromtheuserenvironment.Thecapturedframes areprocessedusingcomputervisiontechniquestodetecthandmovements.ThesystemutilizesanAI-basedhandtracking modeltoextractkeyhandlandmarksandanalyzegesturepatternscontinuously.

Eachframeisprocessedtoidentifythepositionandmovementoffingersusinglandmarkdetection.Theselandmarksareused torecognizepredefinedgestures.Adual-stagegesturedetectionmechanismisimplemented,wherethesystemfirstdetectsa closedfistandthenidentifiesathumbfoldtoconfirmadistresssignal.

Theprocessingunitcontinuouslyreceivesinputfromthewebcamandconvertsthevisualdataintomeaningfulinformation using image processing techniques. The system analyzes this data in real time and compares the detected gestures with predefined conditions programmed in the system. If the detected gestures match the predefined sequence, the system interpretsitasanemergencycondition.Otherwise,thesystemcontinuesmonitoringwithoutinterruption.

Thesystemworkflowisorganizedintofourmainstages:videocapture,gestureprocessing,validation,andalertgeneration. First,thewebcamcapturesreal-timevideoframes.Second,thesystemprocessestheframesanddetectshandlandmarks. Finally, if a valid gestureis confirmed, the system generates an alert and sends notifications through the communication module.

2.2 Gesture Detection and Monitoring

Thegesturedetectionmodulecontinuouslymonitorshandmovementstoidentifypredefineddistressgestures.Thesystem usesawebcamtocapturelivevideoandprocesseseachframeusingcomputervisionandAI-basedhandtrackingtechniques.

TheMediaPipeframeworkisusedtodetect21handlandmarkpoints,whichrepresentthepositionoffingersandjoints.These landmarksareanalyzedtodeterminethestateofeachfingerandidentifyspecificgesturepatterns.

Aclosedfistgestureisdetectedbyanalyzingthepositionofallfingers,whilethethumbfoldisidentifiedbycomparingthe thumbpositionwithitsjoint.Thesystemcontinuouslytracksthesemovementsandvalidatesthegesturesequence.

Theprocesseddataisanalyzedinrealtime,andifthegesturesequenceexceedspredefinedconditions,thesystemtriggersan alert.Continuousmonitoringensuresaccuratedetectionandimprovesthereliabilityofthesystem

International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395-0056

Volume: 13 Issue: 03 | mar 2026 www.irjet.net p-ISSN: 2395-0072

Thegesturevalidationmoduleisresponsibleforconfirmingwhetherthedetectedhandmovementsrepresentavaliddistress signal.Thisisachievedusingapredefinedsequence-basedgesturerecognitionapproachintegratedwiththesystem.

Thesystemusesadual-stagevalidationmechanismtoimproveaccuracy.Initially,aclosedfistgestureisdetectedusinghand landmarkanalysis.Oncethisconditionissatisfied,thesystemwaitsforthesecondgesture,whichinvolvesfoldingthethumb inward.Thissequentialdetectionensuresthataccidentalhandmovementsdonottriggerfalsealerts.

To avoid false detection caused by random gestures or noise, the system performs condition checking and frame-based validation before confirming the gesture. Only when both stages are completed within a specific time frame, the system interpretsitasavalidemergencysignal.

Furthermore,thevalidatedgestureinformationisimmediatelyprocessedbythesystemforalertgeneration.Thisvalidation mechanismimprovesreliabilityandensuresthatonlyintentionalgesturesactivatetheemergencyresponse,enhancingoverall systemperformance.

Volume: 13 Issue: 03 | mar 2026 www.irjet.net p-ISSN: 2395-0072

2.4 Processing Unit

Theprocessingunitservesasthecentralcomponentofthesystem,responsibleforanalyzingvideoinputandrecognizing gestures in real time. It receives video frames from the webcam and processes them using computer vision and artificial intelligencetechniques.

Thesystemextractshandlandmarksfromeachframeusingapre-trainedhandtrackingmodel.Theselandmarksarethen analyzed to determine finger positions and gesture patterns. The processed data is compared with predefined gesture conditionsprogrammedintothesystem.

Theprocessingunitcontinuouslyevaluatestheincomingdatatoidentifywhetherthegesturematchestherequiredsequence. Ifthegestureconditionsaresatisfied,thesystemconfirmsthepresenceofadistresssignal.Otherwise,thesystemcontinues monitoring without interruption. In addition, the processing unit manages the execution of the alert mechanism by coordinatingbetweengesturedetectionandcommunicationmodules.Itsreal-timeprocessingcapabilityensuresfastresponse andaccuratedetection,makingitreliableforsafetyapplications.

2.5 Alert Communication and Data Transmission

The communication module enables real-time transmission of emergency alerts to predefined contacts. The system uses internet-basedcommunicationprotocolstosendnotificationswhenavaliddistressgestureisdetected.Onceanemergency gesture is confirmed, the system generates an alert message and sends it through an email service using SMTP protocol, containingessentialinformationindicatingthattheuserisindistress.

The system ensures that the alert is delivered immediately, enabling quick response from the receiver, and operates continuouslywhilebeingactivatedonlyduringemergencyconditions.Byintegratinggesturerecognitionwithautomatedalert transmission, the system provides a reliable and efficient safety solution, improving response time and ensuring timely assistanceduringcriticalsituations.

2.6 Alert and Notification System

Thealertsystemensuresthatusersreceivetimelyinformationduringemergencysituations.Whenavaliddistressgestureis detected, the system generates an immediate alert notification. Once the gesture is confirmed, an emergency message is automaticallysenttopredefinedcontactsthroughinternet-basedcommunication,deliveredviaemailservicesandincludinga

International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395-0056

Volume: 13 Issue: 03 | mar 2026 www.irjet.net p-ISSN: 2395-0072

distress message indicating that the user requires help. This alert mechanism improves safety by enabling quick communicationandfasteremergencyresponse.

2.7 System Integration and Operation

The entire system operates through the integration of video input, gesture recognition, processing modules, and communicationinterfaces,whereeachcomponentworkstogethertoensureefficientmonitoring,analysis,andtransmissionof emergency alerts, enabling real-time detection of distress gestures and rapid alert generation. Initially, the webcam continuouslycapturesreal-timevideodatafromtheuserenvironment,andthecapturedframesareprocessedusingcomputer visiontechniquestodetecthandmovementswhileextractinghandlandmarkpointsusinganAI-basedmodelthatrepresents thepositionandorientationoffingers.

Theprocesseddataisthensenttothesystem'sprocessingunit,wheregesturerecognitionisperformedbyanalyzinghand landmarksandcomparingthemwithpredefinedgestureconditions,applyingadual-stagevalidationprocesstoconfirmthe distressgesturesequence.

Ifthedetectedgesturesmatchthepredefinedconditions,thesystemidentifiesitasanemergencyandactivatesthealert mechanism;otherwise,itcontinuesmonitoringwithoutinterruption.Onceavalidgestureisconfirmed,thecommunication moduletransmitsthealertmessagetopredefinedcontactsusinginternet-basedprotocols,ensuringinstantdeliveryandquick responsefromthereceiver.

The integrated operation of gesture detection and alert communication enhances system reliability and ensures timely emergencyresponse,providingacontactless,efficient,andreal-timesafetysolutionforusers.

3. WORKING

Theproposedsystemworksbycontinuouslymonitoringhandgesturesusingawebcamanddetectingdistresssignalsthrough gesturerecognitiontechniques.Thesystemcapturesreal-timevideoinputandprocessesittoidentifypredefinedgesture patterns.

Initially,thewebcamrecordsvideoframes,whichareanalyzedusingcomputervisionalgorithms.Thesystemdetectshand landmarksandtracksfingermovementstorecognizegestures.Theextracteddataisprocessedbythesystem,whichcompares thedetectedgestureswithpredefinedconditions.Ifthegesturesmatchtherequiredsequence,thesystemidentifiesitasan emergencycondition.

Ifnovalidgestureisdetected,thesystemcontinuesnormalmonitoring.However,whenadistressgestureisconfirmed,the systemactivatesthealertmechanism.Oncetriggered,thesystemsendsanemergencynotificationtopredefinedcontactsusing internet-basedcommunication.

International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395-0056

Volume: 13 Issue: 03 | mar 2026 www.irjet.net p-ISSN: 2395-0072

4. FLOW CHART

Thesystemstartsbyinitializingthewebcamandprocessingmodules.Afterinitialization,thesystemcontinuouslycaptures videoframesanddetectshandlandmarks.Thesystemprocessesthegesturedataandcomparesitwithpredefinedconditions. Ifthegesturedoesnotmatch,thesystemcontinuesmonitoring.

International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395-0056

Volume: 13 Issue: 03 | mar 2026 www.irjet.net p-ISSN: 2395-0072

5. CONCLUSIONS

The HERSAFETY system presents an effective and reliable solution for women's safety through intelligent, contactless emergencycommunication.Byleveragingcomputervision,MediaPipehandtracking,andSMTP-basedalerttransmission,the systemenablessilentdistresssignalingwithouttheneedforphysicalinteraction.Thedual-stagegesturevalidationmechanism significantlyreducesfalsealarms,whilereal-timeprocessingensuresrapidresponse.ThesystemisimplementedusingPython, OpenCV, and Media Pipe, making it lightweight and accessible without specialized hardware. The proposed solution demonstrateshighdetectionaccuracyandlowresponsetime,establishingitasapracticalandscalableapproachtopersonal safetyincriticalsituations.

ACKNOWLEDGEMENT

TheauthorswouldliketothanktheDepartmentofComputerScienceandEngineering,VivekanandhaCollegeofEngineering forWomen,TamilNadu,India,fortheirguidanceandsupportthroughoutthisproject.

REFERENCES

[1]S.S.RautarayandA.Agrawal,"VisionBasedHandGestureRecognitionforHumanComputerInteraction:ASurvey,"in ArtificialIntelligenceReview,vol.43,no.1,pp.1-54,2015.

[2]MediaPipeTeam,"MediaPipeHands:On-DeviceReal-TimeHandTracking,"GoogleResearch,TechnicalReport,2023.

[3]G.Bradski,"TheOpenCVLibrary,"inDr.Dobb'sJournalofSoftwareTools,vol.25,no.11,pp.120-125,2000.

[4]Z.Zhang,"MicrosoftKinectSensorandItsEffect,"inIEEEMultimedia,vol.19,no.2,pp.4-10,2012.

[5]R.Szeliski,ComputerVision:AlgorithmsandApplications,Springer,NewYork,2010.

[6]A.KulshreshthaandR.Sharma,"Real-TimeHandGestureRecognitionforEmergencyAlertSystemsUsingComputerVision," inInternationalJournalofComputerVisionandImageProcessing,vol.12,no.3,pp.45-58,2022.

[7]P.Nair,S.Menon,andV.Krishnan,"AI-BasedContactlessSafetySystemsforWomenUsingGestureRecognitionandIoT Alerts,"inJournalofArtificialIntelligenceandSafetyTechnologies,vol.5,no.2,pp.78-91,2023.

[8]M.DasandT.Roy,"MediaPipe-BasedHandLandmarkDetectionforAssistiveandSafetyApplications,"inInternational JournalofMachineLearningandEmbeddedSystems,vol.8,no.1,pp.33-47,2022.

[9] R. Gupta and A. Singh, "Automated Emergency Notification Systems Using Python and SMTP Protocol for Real-Time PersonalSafety,"inJournalofEmbeddedSystemsandIoTApplications,vol.6,no.4,pp.112-125,2021.

[10] K. Patel and S. Joshi, "Gesture-Based Human-Computer Interaction for Accessibility and Assistive Technology Applications,"inIEEETransactionsonHuman-MachineSystems,vol.51,no.3,pp.220-234,2021.

[11]A.Verma,N.Sharma,andD.Mehta,"SilentEmergencyAlertingUsingComputerVisionforWomenSafetyinPublicSpaces," inInternationalConferenceonAIandWomen'sSafety,pp.89-97,2023.

[12]B.ChenandY.Liu,"DeepLearningApproachesforReal-TimeHandPoseEstimationandGestureClassification,"inPattern RecognitionLetters,vol.145,pp.56-64,2021.