Show simple item record

dc.contributor.advisorChakrabarty, Amitabha
dc.contributor.authorIrtiza, Md. Samin
dc.contributor.authorAhmed, Fattah
dc.contributor.authorHaque, Md. Tahmidul
dc.contributor.authorTamim, Arifur Rahman
dc.contributor.authorSultana, Samia
dc.date.accessioned2023-10-16T07:43:04Z
dc.date.available2023-10-16T07:43:04Z
dc.date.copyright©2022
dc.date.issued2022-09-28
dc.identifier.otherID 18101429
dc.identifier.otherID 18101442
dc.identifier.otherID 18101570
dc.identifier.otherID 18101510
dc.identifier.otherID 18101446
dc.identifier.urihttp://hdl.handle.net/10361/21848
dc.descriptionThis thesis is submitted in partial fulfillment of the requirements for the degree of Bachelor of Science in Computer Science, 2022.en_US
dc.descriptionCataloged from PDF version of thesis.
dc.descriptionIncludes bibliographical references (pages 31-33).
dc.description.abstractIt is undeniable that in recent years, exceptional progress has been made toward building the most accurate and efficient object detectors. However, existing low- light object detectors still require a substantial amount of resources to perform at their best. Our main goal in this research is to train and evaluate recently developed deep learning object detection models on low-light images and see if they can show decent performance without any additional enhancement networks. Furthermore, we aim to achieve those results with minimum computational cost. In this research, we have created our own custom dataset from a publicly available insect image dataset called ‘IP102’. The new dataset now named ‘IP013’ consists of 13 classes of insects and approximately 8k annotated images. Moreover, we chose recently developed YOLOv7 and DETR object detectors and compared their performance against now older state-of-the-art RetinaNet and EfficientDet deep learning models. YOLOv7, EfficientDet, and RetinaNet are purely CNN-based models whereas DETR uses a Transformer as both encoder and decoder and a CNN as the backbone. Our research shows that YOLOv7 outperforms all of the other models with a mAP0.5:.95 of 45.9 while using the lowest training time and the model that used the least computational resources was EfficientDet which admittedly showed lackluster mAP0.5:.95 of 33.2 with only 3.9M parameters and using 2.5 GFLOPs.en_US
dc.description.statementofresponsibilityMd. Samin Irtiza
dc.description.statementofresponsibilityFattah Ahmed
dc.description.statementofresponsibilityMd. Tahmidul Haque
dc.description.statementofresponsibilityArifur Rahman Tamim
dc.description.statementofresponsibilitySamia Sultana
dc.format.extent43 pages
dc.language.isoenen_US
dc.publisherBrac Universityen_US
dc.rightsBrac University theses are protected by copyright. They may be viewed from this source for any purpose, but reproduction or distribution in any format is prohibited without written permission.
dc.subjectLow-light imagesen_US
dc.subjectEnhancement networken_US
dc.subjectDeep learningen_US
dc.subjectObject detectionen_US
dc.subjectInsecten_US
dc.subject.lcshMachine learning
dc.subject.lcshComputer vision
dc.titleAnalysis of deep learning models on low-light pest detectionen_US
dc.typeThesisen_US
dc.contributor.departmentDepartment of Computer Science and Engineering, Brac University
dc.description.degreeB.Sc. in Computer Science


Files in this item

Thumbnail

This item appears in the following Collection(s)

Show simple item record