Ciencias Exactas y Ciencias de la Salud
Permanent URI for this collectionhttps://hdl.handle.net/11285/551014
Pertenecen a esta colección Tesis y Trabajos de grado de los Doctorados correspondientes a las Escuelas de Ingeniería y Ciencias así como a Medicina y Ciencias de la Salud.
Browse
Search Results
- A methodology to select downsized object detection algorithms for resource-constrained hardware using custom-trained datasets(Instituto Tecnológico y de Estudios Superiores de Monterrey, 2025-12-03) Medina Rosales, Adán; Ponce Cruz, Pedro; emipsanchez; López Cadena, Edgar Omar; Montesinos Silva, Luis Arturo; Balderas Silva, David Christopher; Ponce Espinosa, Hiram Eredín; School of Engineering and Sciences; Campus Ciudad de MéxicoDownsized object detection algorithms have gained relevance with the exploration of edge computing and implementation of these algorithms in small mobile devices like drones or small robots. This has led to an exponential growth of the field with several new algorithms being presented every year. With no time to test them all most benchmark focus on testing the full sized versions and comparing training results. This however, creates a gap in the state of the art since no comparisons of downsized algorithms are being presented, specifically using custom built datasets to train the algorithms and restrained hardware devices to implement them. This work aims to provide the reader with a comprehensive understanding of several metrics obtained not only from training metrics, but also from implementation to have a more complete picture on the behavior of the downsized algorithms (mostly from the YOLO algorithm family), when trained with small datasets, by using a fiber extrusion device with three classes: one that has no defects, one that is very similar looking with small changes and one that has a more immediate tell in the difference, showcasing how good the algorithms tell apart each class using two different size of datasets, while also providing information on training times and different restrained hardware implementation results. Providing results on implementation metrics as well as training metrics.
- Real-time armed individual detection in video surveillance usingdeep learning and heuristic approaches(Instituto Tecnológico y de Estudios Superiores de Monterrey, 2024-12) Amado Garfias, Alonso Javier; Conant Pablos, Santiago Enrique; emipsanchez; Ortiz bayliss, José Carlos; Tarashima Marín Hugo; Gutiérrez Rodríguez, Andrés Eduardo; School of Engineering and Sciences; Campus MonterreyThis researchaimstoenhancetheautomaticidentificationofarmedindividualsinvideo surveillanceinreal-time.Theproposedmethodologyinvolvesthedevelopmentofalgorithms specifically designedforthedetectionofindividualscarryinghandguns,whichincludepistols and revolvers.Toachievethis,theYOLOv4modelhasbeenselectedtodetectindividuals, handguns, andfaces.Subsequently,real-timeinformationisextractedfromtheYOLOmodel, including boundingboxcoordinates,distances,andintersectionareasbetweenhandgunsand individualswithineachvideoframe.Thisinformationfeedsourheuristicsanddifferentma- chine learning(ML)proposed,facilitatingtherecognitionofarmedindividuals.Severalchal- lenges mustbeaddressed,suchasocclusion,concealedguns,andproximityofindividualsto one another.Itencouragesthedevelopmentandcomparisonofdifferenttypesofsolutions. Theyaremadeupofthreeheuristics,seven-armedpeopledetectors(APD),and44APDto use ineachvideoframe(APD4F). The heuristicsaretheDeterministicMethodofCenters(DMC),theDeterministicMethod of Distances(DMD),andtheDeterministicMethodofIntersections(DMI).Furthermore, the APDmodelsareRandomForestClassifier(RFC-APD),MultilayerPerceptron(MLP- APD), k-Nearest-Neighbors(KNN-APD),SupportVectorMachine(SVM-APD),Logistic Regression(LR-APD),NaiveBayes(NB-APD),andGradientBoostingClassifier(GBC- APD). Thereby,IproposetocreateselectorsfordecidingwhichAPDtouseineachvideo frame (APD4F)toimprovethedetectionresults.Besides,weimplementedtwotypesof APD4Fs, onebasedonaRandomForestClassifier(RFC-APD4F)andanotherinaMultilayer Perceptron (MLP-APD4F).Wedeveloped44APD4FscombiningsubsetsofsixAPDs.The most ofAPD4FoutperformedoftheindependentuseofallAPDs.Amultilayerperceptron- based APD4F,whichcombinesanMLP-APD,aNB-APD,andaLR-APD,presentedthebest performance, achievinganaccuracyof95.84%,arecallof99.28%andanF1scoreof96.07%. This researchalsoproposesasolutiontooptimizetheproblemofdetectingarmedpeople when theweaponisnotvisible.Therefore,weapplyrecurrentneuralnetworks,suchasLong Short TermMemory(LSTM),topredictthecoordinatesoftheguns.Inthisway,itispossible to haveapredictionofarmedpeopleatalltimes.ThemeasurementbetweentheYOLO handgun detectionboundingboxesandtheLSTMpredictionresultedinanIoUof65.23%. When thefirearmdetectionbytheobjectdetectorisinterrupted,theweapon’spositionis generated bytheLSTMmodelsthat,togetherwiththeAPDs,identifythearmedpeople. When theLSTMsdeliveredtheirpredictionstotheAPDs,theNB-APDdemonstratedthe best performance,achievinganaccuracyof80.93%.TheLSTMsallowedtheanalysisof 5,288 recordsofthetestvideothatcouldnotbeanalyzedbeforeduetothelackofknowledge of thegun’sposition.